{ "paper_id": "O13-1017", "header": { "generated_with": "S2ORC 1.0.0", "date_generated": "2023-01-19T08:03:46.829548Z" }, "title": "", "authors": [ { "first": "Chang-Hong", "middle": [], "last": "\u6797\u6636\u5b8f", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Central University", "location": {} }, "email": "" }, { "first": "", "middle": [], "last": "Lin", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Central University", "location": {} }, "email": "" }, { "first": "Ernestasia", "middle": [], "last": "\u897f\u96c5\u6069", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Central University", "location": {} }, "email": "" }, { "first": "", "middle": [], "last": "Siahaan", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Central University", "location": {} }, "email": "" }, { "first": "Bo-Wei", "middle": [], "last": "\u9673\u4f2f\u7152", "suffix": "", "affiliation": { "laboratory": "", "institution": "Cheng Kung University", "location": {} }, "email": "" }, { "first": "", "middle": [], "last": "Chen", "suffix": "", "affiliation": { "laboratory": "", "institution": "Cheng Kung University", "location": {} }, "email": "" }, { "first": "Hsiang-Lung", "middle": [], "last": "\u838a\u7965\u74cf", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Central University", "location": {} }, "email": "" }, { "first": "", "middle": [], "last": "Chuang", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Central University", "location": {} }, "email": "" }, { "first": "Wen-Chi", "middle": [], "last": "\u8b1d\u73f3\u68cb", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Central University", "location": {} }, "email": "" }, { "first": "", "middle": [], "last": "Hsieh", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Central University", "location": {} }, "email": "" }, { "first": "Jia-Ching", "middle": [], "last": "\u738b\u5bb6\u6176", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Central University", "location": {} }, "email": "" }, { "first": "", "middle": [], "last": "Wang", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Central University", "location": {} }, "email": "" } ], "year": "", "venue": null, "identifiers": {}, "abstract": "", "pdf_parse": { "paper_id": "O13-1017", "_pdf_hash": "", "abstract": [], "body_text": [ { "text": "EQUATION", "cite_spans": [], "ref_spans": [], "eq_spans": [ { "start": 0, "end": 8, "text": "EQUATION", "ref_id": "EQREF", "raw_str": "\u2212 = \u2212 = N k L l k qk q l t s l h t x 1 1 0 ) ( ) ( ) ( (1) \u5176\u4e2d x q \u662f\u611f\u6e2c\u5668 q \u5c0d\u61c9\u7684\u6df7\u5408\u8a0a\u865f\uff0cs k \u70ba\u6e90\u8a0a\u865f k\uff0ch qk \u5247\u662f\u8a9e\u8005 k \u5230\u9ea5\u514b\u98a8 q \u7684\u8108\u885d\u97ff \u61c9\uff0c\u4e26\u4e14\u4ee4\u9019\u500b\u6ffe\u6ce2\u5668 (Filter) \u7684\u578b\u5f0f\u70ba\u4e00\u500b L \u968e (L-Tap) \u7684\u6709\u9650\u8108\u885d\u97ff\u61c9 (Finite Impulse Response, FIR)\u6ffe\u6ce2\u5668\u3002\u7531\u65bc\u8a9e\u97f3\u5728\u6642\u9593\u57df\u4e0a\u7684\u7a00\u758f\u7279\u6027\u4e26\u4e0d\u660e\u986f\uff0c\u6240\u4ee5\u6211\u5011\u63a1\u7528\u77ed\u6642 \u5085\u5229\u8449\u8f49\u63db(Short Time Fourier Transform, STFT) \uff0c\u4ee5\u53d6\u6a23\u983b\u7387 f s \u5c07\u6642\u9593\u57df\u4e0a\u7684\u6df7\u5408\u8a0a \u865f x q (t)\u8f49\u63db\u6210\u983b\u7387\u57df\u4e0a\u7684\u6642\u9593\u5e8f\u5217 x q (f,\u03c4)\uff0c\u5176\u4e2d f \u662f\u67d0\u500b\u983b\u5e36\uff0c\u03c4 \u70ba\u77ed\u6642\u5085\u5229\u8449\u8f49\u63db\u97f3\u7a97 \u7684\u6307\u6a19(Frame Index) \u3002\u5728\u6642\u983b\u57df\u4e0a\u57f7\u884c\u76f2\u8a0a\u865f\u6e90\u5206\u96e2\u7684\u53e6\u4e00\u500b\u597d\u8655\u662f\u6211\u5011\u53ef\u4ee5\u5c07\u65cb\u7a4d \u6df7\u5408\u904e\u7a0b\u55ae\u7d14\u8996\u70ba\u5404\u500b\u983b\u5e36\u7684\u77ac\u6642\u6df7\u5408\u578b\u5f0f\uff0c\u5373\u5982\u540c\u4ee5\u4e0b\u4e4b\u6558\u8ff0\u3002 N M N M N k k k C f H C f S C f X f S f H f S f H f X \u00d7 \u00d7 \u00d7 = \u2208 \u2208 \u2208 = = \u2211 ) ( , ) , ( , ) , ( where , ) , ( ) ( ) , ( ) ( ) , ( 1 1 1 \u03c4 \u03c4 \u03c4 \u03c4 \u03c4 (2) \u5176\u4e2d X(f,\u03c4)\u548c S(f,\u03c4)\u5206\u5225\u4ee3\u8868\u6df7\u5408\u8a0a\u865f\u4ee5\u53ca\u4f86\u6e90\u8a0a\u865f\u5728\u6642\u983b\u57df\u4e0a\u7684\u6210\u4efd\u3002H(f)\u5247\u662f\u67d0\u4e00\u500b \u983b\u5e36\u7684\u6df7\u5408\u77e9\u9663\u3002\u7136\u800c\uff0c\u5047\u8a2d\u5728\u4e00\u500b\u6642\u983b\u9ede\u4e0a\uff0c\u53ea\u6709\u4e00\u500b\u4f86\u6e90\u8a0a\u865f\u5728\u6d3b\u52d5\uff0c\u6211\u5011\u4ee4 H k (f) \u662f H(f)\u7684\u7b2c k \u500b\u884c\u5411\u91cf\uff0c\u5247\u53ef\u5c07\u5f0f\u5b50(2)\u7c21\u5316\u70ba\uff1a { } N k f S f H f X k k , , 1 , ) , ( ) ( ) , ( L \u2208 = \u03c4 \u03c4 (3) \u6240\u8b02\u7684\u6ce2\u675f\u5f62\u6210(Beamforming) \uff0c\u5373\u70ba\u4e00\u7a2e\u7a7a\u9593\u4e0a\u4e4b\u6ffe\u6ce2\u5668\uff0c\u5b83\u5229\u7528\u8a0a\u865f\u7684\u7a7a\u9593\u95dc \u4fc2\uff0c\u5e0c\u671b\u80fd\u5920\u5c0d\u4e0d\u540c\u65b9\u5411\u7684\u8a0a\u865f\u505a\u51fa\u4e0d\u540c\u7684\u589e\u76ca\uff0c\u4ee5\u9054\u5230\u7a7a\u9593\u6ffe\u6ce2\u7684\u6548\u679c\uff0c\u85c9\u4ee5\u5206\u96e2\u7a7a \u9593\u4e2d\u4e0d\u540c\u65b9\u5411\u8072\u6e90\u7684\u8a0a\u865f\u3002\u4f9d\u6ce2\u675f\u5f62\u6210\u5b9a\u7406\uff0c\u6211\u5011\u9760\u8457\u9ea5\u514b\u98a8\u9663\u5217\u7684\u4f86\u6e90\u8a0a\u865f\u65b9\u5411\u548c\u6642 \u9593\u5ef6\u9072\u53bb\u8fd1\u4f3c\u6df7\u5408\u904e\u7a0b\u3002\u56e0\u6b64\u7576\u983b\u7387\u70ba f \u6642\uff0c\u8a9e\u8005 k \u5230\u9ea5\u514b\u98a8 q \u7684\u6df7\u5408\u4fc2\u6578\u53ef\u8868\u793a\u70ba\uff1a k q d fc j qk qk e g f h \u03b8 \u03c0 cos 2 1 ) ( \u2212 = (4) \u5176\u4e2d g qk \u70ba\u8a0a\u865f k \u81f3\u9ea5\u514b\u98a8 q \u7684\u589e\u76ca\u503c\uff0cd q \u8868\u611f\u6e2c\u5668 q \u8207\u9ea5\u514b\u98a8\u9663\u5217\u4e2d\u5fc3\u4e4b\u9593\u7684\u8ddd\u96e2\uff0c\u03b8 k \u662f\u6e90\u8a0a\u865f k \u5c0d\u61c9\u5230\u9ea5\u514b\u98a8\u9663\u5217\u7684\u89d2\u5ea6\u3002\u6211\u5011\u53ef\u5229\u7528\u5f0f\u5b50(4)\uff0c\u5c07\u6df7\u5408\u77e9\u9663\u8868\u73fe\u6210\u4e0b\u9762\u7684 \u5f62\u5f0f\uff0c\u5f80\u5f8c\u6709\u95dc\u6df7\u5408\u77e9\u9663\u7684\u63a8\u5c0e\u904e\u7a0b\uff0c\u591a\u6578\u90fd\u662f\u5efa\u7acb\u5728\u9019\u500b\u9810\u8a2d\u5f62\u5f0f\u4e4b\u4e0a\u3002 \u23a5 \u23a5 \u23a5 \u23a6 \u23a4 \u23a2 \u23a2 \u23a2 \u23a3 \u23a1 = ) ( ) ( ) ( ) ( ) ( 1 1 11 f h f h f h f h f H MN M N L M O M L (5) (\u4e8c) \u3001\u7279\u5fb5\u503c\u9078\u53d6 \u6211\u5011\u5b9a\u7fa9\u4e86\u5169\u500b\u6df7\u5408\u8a0a\u865f\u7684\u7279\u5fb5\u53c3\u6578(Level-Ratio \u548c Phase-Difference)[11]\u3002\u5229\u7528 \u89c0\u5bdf\u8cc7\u6599\u7684\u4e8c\u968e\u7bc4\u6578\u5c0d\u6df7\u5408\u8a0a\u865f\u7684\u7d55\u5c0d\u503c\u983b\u8b5c(Magnitude Spectrum)\u505a\u6b63\u898f\u5316\uff0c\u6211\u5011 \u7a31\u4e4b\u70ba Level-Ratio\uff0c\u9019\u908a\u7528 ) , ( \u03c4 \u03c8 f L q ( \u8868\u793a\uff1b\u81f3\u65bc Phase-Difference \u88ab\u5b9a\u7fa9\u6210\u8207\u4e00\u500b\u6307\u5b9a\u7684 \u6df7\u5408\u8a0a\u865f\u4e4b\u9593\u7684\u76f8\u4f4d\u89d2\u5ea6\u5dee\uff0c\u4ee5 ) ,\u03c4 \u03c8 f P q \u4f86\u8868\u793a\u3002\u5b83\u5011\u7684\u8868\u793a\u5f0f\u5206\u5225\u986f\u793a\u5982\u4e0b\uff1a 2 ) , ( ) , ( ) , ( \u03c4 \u03c4 \u03c4 \u03c8 f X f x f q L q = (6) [ ] [ ] ) , ( ) , ( ) , ( 1 \u03c4 \u03c6 \u03c4 \u03c6 \u03c4 \u03c8 f x f x f q P q \u2212 = (7) \u5176\u4e2d\u03c6\u70ba\u76f8\u4f4d\u7684\u904b\u7b97\u5b50\u3002\u7136\u5f8c\u5229\u7528\u4e00\u500b\u8907\u6578\u8868\u793a\u5f0f(Complex Representation)\u4f86\u8868\u73fe\u9019 \u5169\u500b\u7279\u5fb5\u53c3\u6578\u3002 )] , ( exp[ ) , ( ) , ( \u03c4 \u03c8 \u03c4 \u03c8 \u03c4 \u03c8 f j f f P q L q q \u00d7 = (8) \u65bc \u662f \u6211 \u5011 \u5f97 \u5230 \u4e86 \u4e00 \u500b \u65b0 \u7684 \u6a23 \u672c \u578b \u614b (Sample Form) \uff0c \u7531 M \u500b Level-Ratio \u548c Phase-Difference \u7d44\u6210\u7684\u8907\u6578\u503c\u6240\u69cb\u6210\u3002\u5c07\u539f\u5148\u7684\u89c0\u5bdf\u8cc7\u6599\u8f49\u63db\u6210\u9019\u6a23\u7684\u8cc7\u6599\u578b\u5f0f\u5f8c\uff0c \u6211\u5011\u5373\u53ef\u4f7f\u7528\u9019\u4e9b\u65b0\u5efa\u7acb\u7684\u6a23\u672c\uff0c\u505a\u5f8c\u7e8c\u7684\u8655\u7406\u548c\u8a0a\u865f\u5206\u6790\uff0c\u5305\u62ec\u4f30\u8a08\u6e90\u8a0a\u865f\u500b\u6578\u4ee5\u53ca \u6df7\u5408\u77e9\u9663\u3002\u4ee4 T \u70ba\u5411\u91cf\u7684\u8f49\u7f6e\uff0c\u5247\u6a23\u672c\u578b\u614b\u8868\u793a\u5982\u4e0b\uff1a [ ] T M f f f ) , ( ) , ( ) , ( 1 \u03c4 \u03c8 \u03c4 \u03c8 \u03c4 L = \u03a8 (9) (\u4e09) \u3001\u6df7\u5408\u77e9\u9663\u4f30\u6e2c \u5229\u7528\u8457\u540d\u4e14\u61c9\u7528\u5ee3\u6cdb\u7684\u5206\u7fa4\u65b9\u6cd5 K-Means \u6f14\u7b97\u6cd5\uff0c\u5c07\u6a23\u672c\u578b\u614b\u5206\u5272\u5230 N \u500b\u7fa4\u805a C i ,\u2026,C N \u4e2d\uff0c\u4e26\u4e14\u5229\u7528\u4e0b\u9762\u7684\u5f0f\u5b50\u7372\u5f97\u6df7\u5408\u5411\u91cf\uff1a { } N i C h i C i i , , 1 , 1 L \u2208 \u03a8 = \u2211 \u2208 \u03a8 (10) \u5176\u4e2d|C i |\u4ee3\u8868\u7b2c i \u500b\u7fa4\u805a\u64c1\u6709\u7684\u6a23\u672c\u6578\u3002\u7136\u800c\u6bcf\u500b\u6df7\u5408\u5411\u91cf\u90fd\u6703\u5c0d\u61c9\u5230\u4e00\u500b\u4f86\u6e90\u8a0a\u865f\u3002 \u56e0\u70ba\u6211\u5011\u662f\u6839\u64da\u6bcf\u500b\u983b\u5e36\u4e0a\u7684\u6642\u9593\u5e8f\u5217\u53bb\u4f30\u6e2c\u6df7\u5408\u77e9\u9663\uff0c\u6240\u4ee5\u5404\u500b\u983b\u5e36\u57f7\u884c\u904e K-Means \u6f14\u7b97\u6cd5\u5f8c\u90fd\u6703\u56de\u50b3 N \u500b\u7fa4\u805a\uff0c\u4e26\u6c42\u51fa\u4ee3\u8868\u7684 h i \u3002\u6700\u5f8c\uff0c\u5982\u4f55\u78ba\u8a8d h i \u5728\u77e9\u9663\u4e2d\u7684\u4f4d\u7f6e\u4e5f \u662f\u4e00\u500b\u5f88\u91cd\u8981\u7684\u554f\u984c\u3002 \u6839\u64da\u5f0f\u5b50(4)\uff0c\u53ef\u4ee5\u5f97\u77e5 i s r d d fc j si ri i i e g g s h r h \u03b8 \u03c0 cos ) ( 2 1 ) ( ) ( \u2212 \u2212 = (11) \u6240\u4ee5\u7d93\u63a8\u5c0e\u5f8c\uff0cDOA \u53ef\u4ee5\u7531\u4e0b\u5f0f\u7372\u5f97 ) ( 2 ) ( ) ( cos 1 1 s r i i i d d fc s h r h \u2212 \u239f \u23a0 \u239e \u239c \u239d \u239b = \u2212 \u2212 \u03c0 \u03c6 \u03b8 (12) \u5176\u4e2d r\u3001s \u662f\u9ea5\u514b\u98a8\u9663\u5217\u4e2d\u5169\u652f\u8ddd\u96e2\u6700\u8fd1\u7684\uff0c\u5b83\u5011\u5728\u6df7\u5408\u5411\u91cf h i \u4e2d\u6240\u5c0d\u61c9\u5230\u7684\u6307\u6a19\uff1bd \u8868\u793a r\u3001s \u5169\u652f\u9ea5\u514b\u98a8\u4e4b\u9593\u7684\u8ddd\u96e2\u3002\u56e0\u70ba\u6211\u5011\u5c0d\u6240\u6709 h i (i=1,\u2026,N)\u90fd\u5075\u6e2c DOA\uff0c\u6240\u4ee5\u5171\u5f97 \u5230\u4e86 N \u500b\u89d2\u5ea6\u503c\u3002\u6700\u5f8c\u6839\u64da\u9019\u500b\u7d50\u679c\u78ba\u5b9a h i \u5728\u6df7\u5408\u77e9\u9663\u4e2d\u6240\u5c0d\u61c9\u7684\u884c\u7d22\u5f15\u3002 \u5047\u8a2d\u6df7\u5408\u8a0a\u865f\u7684\u67d0\u500b\u6642\u983b\u9ede X(f,\u03c4)\uff0c\u53ea\u6709\u6e90\u8a0a\u865f k \u70ba\u975e\u96f6\u7684\u503c\u3002\u900f\u904e\u5f0f\u5b50(3)\u548c\u5f0f\u5b50 (4)\uff0c\u53ef\u5c07 X(f,\u03c4)\u53ef\u8868\u73fe\u70ba\uff1a [ ] [ ] [ ] \u23a5 \u23a5 \u23a5 \u23a5 \u23a6 \u23a4 \u23a2 \u23a2 \u23a2 \u23a2 \u23a3 \u23a1 \u00d7 \u00d7 = \u00d7 \u23a5 \u23a5 \u23a5 \u23a5 \u23a6 \u23a4 \u23a2 \u23a2 \u23a2 \u23a2 \u23a3 \u23a1 = + + \u2212 \u2212 \u2212 \u2212 ) ) , ( cos 2 ( ) , ( ) ) , ( cos 2 ( ) ,", "eq_num": "( 1 )" } ], "section": "", "sec_num": null }, { "text": "EQUATION", "cite_spans": [], "ref_spans": [], "eq_spans": [ { "start": 0, "end": 8, "text": "EQUATION", "ref_id": "EQREF", "raw_str": "\u7279\u5fb5\u53c3\u6578\uff0c ) , ( \u03c4 \u03c8 f L q \u548c ) , ( \u03c4 \u03c8 f P q \u5206\u5225\u70ba\uff1a [ ] ) ( ) , ( 1 T Mk k qk L q g g norm g f L = \u03c4 \u03c8", "eq_num": "(14)" } ], "section": "", "sec_num": null }, { "text": "\u5176\u4e2d\uff0c\u7b2c\u4e00\u9805\u70ba\u4e00\u5be6\u6578\u503c\u3002\u85c9\u7531\u4e0a\u5f0f\uff0c\u6211\u5011\u53ef\u4ee5\u8aaa\uff0c\u7576\u8a9e\u97f3\u5177\u6709\u6975\u5ea6\u7a00\u758f\u7684\u6027\u8cea\u6642\uff0c\u53ea \u6703\u56e0\u70ba\u4e3b\u5c0e\u7684\u6e90\u8a0a\u865f\u4e0d\u540c\u9020\u6210 \u03b8 k \u7684\u6539\u8b8a\u800c\u7522\u751f N \u7a2e\u578b\u5f0f\u7684 \u03c8(f,\u03c4)\u3002\u6240\u4ee5\u7576\u7d50\u675f\u5206\u7fa4\u6f14\u7b97 \u6cd5\u4f30\u8a08\u6df7\u5408\u77e9\u9663\u4e4b\u884c\u5411\u91cf\u7684\u7a0b\u5e8f\uff0c\u4e26\u4e14\u89e3\u6c7a\u4e86\u6392\u5217\u554f\u984c\u5f8c\uff0c\u5728\u6700\u7406\u60f3\u7684\u60c5\u6cc1\u4e0b\uff0c\u4e5f\u5c31\u662f \u7576\u6975\u5ea6\u7a00\u758f\u7684\u689d\u4ef6\u6210\u7acb\u6642\uff0c\u6df7\u5408\u77e9\u9663\u6703\u9577\u6210\uff1a \u23a5 \u23a5 \u23a5 \u23a5 \u23a5 \u23a6 \u23a4 \u23a2 \u23a2 \u23a2 \u23a2 \u23a2 \u23a3 \u23a1 \u2212 \u2212 \u2212 \u2212 \u2212 \u2212 \u2212 \u2212 N M M N d d fc j L MN d d fc j L M d d fc j L N d d fc j L L N", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "EQUATION", "cite_spans": [], "ref_spans": [], "eq_spans": [ { "start": 0, "end": 8, "text": "EQUATION", "ref_id": "EQREF", "raw_str": "1 1 1 1 1 1 2 1 1 1 2 1 L M O M L L (17) \u76f2\u8a0a\u865f\u6e90\u5206\u96e2\u5728\u6b20\u5b9a\u7684\u689d\u4ef6\u4e0b\uff0c\u6839\u64da\u5f0f\u5b50(2)\uff0c\u6e90\u8a0a\u865f S \u53ef\u4ee5\u6709\u7121\u9650\u591a\u500b\u89e3\uff0c\u6240\u4ee5 \u6211\u5011\u5229\u7528\u6700\u5c0f\u5316l1 \u7684\u7bc4\u6578\u4ee5\u53ca X=HS \u4f5c\u70ba\u9650\u5236\u5f0f\uff0c\u6b64\u6700\u4f73\u5316\u554f\u984c\u7684\u89e3\u5373\u70ba\u6240\u6c42\u3002\u5982\u4e0b\u5217 \u5f0f\u5b50\u6240\u793a\uff1a X HS t s N i S i i S = = \u2211 . . , ,", "eq_num": ", 1 ," } ], "section": "", "sec_num": null }, { "text": "( x w\u22c5 ) + b = 0 \u6700\u5927\u5316margin 2/|| w || 2 \uff0c\u5176\u4e2d w \u2208 d \u4e14 b \u2208 \u3002\u6839\u64da\u6c7a\u7b56\u51fd\u6578(Decision Function)\u8cc7\u6599\u9ede x \u88ab\u6a19\u8a18\u6210y \u2208 {1, -1} ) ) (( sign ) ( b x w x f + \u22c5 = (19) \u6211\u5011\u53ef\u4ee5\u5728 SVM \u4e2d\u4f7f\u7528\u6838\u65b9\u6cd5(Kernel Methods) \u3002\u9996\u5148\uff0c\u5c07\u5206\u96e2\u8d85\u5e73\u9762\u51fd\u6578\u8868\u793a \u6210\u8cc7\u6599\u9ede x \u7684\u5167\u7a4d\uff0c\u5247\u6c7a\u7b56\u51fd\u6578\u53ef\u4ee5\u5beb\u6210\u4e0b\u9762\u9019\u500b\u5f0f\u5b50\uff1a ) ( sign ) ( 1 b x x x f m i i i + \u22c5 = \u2211 = \u03b1 (20) \u5176\u4e2d\u03b1 \u70ba\u62c9\u666e\u62c9\u65af\u4e58\u6578(Lagrange Multiplier) \uff0ci \u70ba\u5411\u91cf\u7684\u6578\u91cf\u3002\u6b64\u5411\u91cf\u4e58\u7a4d\u53ef\u88ab\u6838\u51fd \u6578(Kernel Function) \u6240\u53d6\u4ee3\uff0c ) , ( i x x k ) ) , ( ( ) ( 1 b x x k x f m i i i + = \u2211 = \u03b1 sign (21) \u85c9\u7531\u4f7f\u7528Mercer's\u7406\u8ad6\uff0c\u6211\u5011\u53ef\u4ee5\u5f15\u5165\u4e00\u500b\u6620\u5c04\u51fd\u6578(Mapping Function)\u03d5( x ) \u4f7f \u5f97 ( k \uff0c ) ( ) ( ) , i j i j x x x x \u03d5 \u03d5 = \u3002\u85c9\u7531\u5c07\u539f\u59cb\u8f38\u5165\u7a7a\u9593 d \u5f71\u5230\u5176\u4ed6\u7a7a\u9593\uff0c\u6b64\u65b9\u6cd5\u63d0\u4f9b\u4e86\u8655\u7406 \u975e\u7dda\u6027\u8cc7\u6599\u7684\u80fd\u529b\u3002 \u6295 (\u4e8c) \u3001\u8072\u97f3\u7279\u5fb5\u503c\u64f7\u53d6 1\u3001\u6885\u723e\u5012\u983b\u8b5c\u4fc2\u6578 \u70ba \u4e86 \u5f9e \u8072 \u97f3 \u8a0a \u865f \u4e2d \u64f7 \u53d6 \u6885 \u723e \u5012 \u983b \u8b5c \u4fc2 \u6578 [", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "Proceedings of the Twenty-Fifth Conference on Computational Linguistics and Speech Processing(ROCLING 2013)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null } ], "back_matter": [], "bib_entries": { "BIBREF0": { "ref_id": "b0", "title": "Audio-visual active speaker racking in cluttered indoor environments", "authors": [ { "first": "F", "middle": [], "last": "Talantzis", "suffix": "" }, { "first": "A", "middle": [], "last": "Pnevmatikakis", "suffix": "" }, { "first": "A", "middle": [ "G" ], "last": "Constantinides", "suffix": "" } ], "year": 2008, "venue": "IEEE Trans. Systems, Man, and Cybernetics", "volume": "38", "issue": "", "pages": "799--807", "other_ids": {}, "num": null, "urls": [], "raw_text": "F. Talantzis, A. Pnevmatikakis, and A. G. Constantinides, \"Audio-visual active speaker racking in cluttered indoor environments,\" IEEE Trans. Systems, Man, and Cybernetics, Part B: Cybernetics, vol. 38, pp. 799-807, Jun. 2008.", "links": null }, "BIBREF1": { "ref_id": "b1", "title": "Versatile recognition using Haar-like feature and cascaded classifier", "authors": [ { "first": "J", "middle": [], "last": "Nishimura", "suffix": "" }, { "first": "T", "middle": [], "last": "Kuroda", "suffix": "" } ], "year": 2010, "venue": "IEEE Sensors Journal", "volume": "10", "issue": "", "pages": "942--951", "other_ids": {}, "num": null, "urls": [], "raw_text": "J. Nishimura and T. Kuroda, \"Versatile recognition using Haar-like feature and cascaded classifier,\" IEEE Sensors Journal, vol. 10, pp. 942-951, May 2010.", "links": null }, "BIBREF2": { "ref_id": "b2", "title": "Wireless sensor networks for home health care", "authors": [ { "first": "C", "middle": [ "R" ], "last": "Baker", "suffix": "" }, { "first": "K", "middle": [], "last": "Armijo", "suffix": "" }, { "first": "S", "middle": [], "last": "Belka", "suffix": "" }, { "first": "M", "middle": [], "last": "Benhabib", "suffix": "" }, { "first": "V", "middle": [], "last": "Bhargava", "suffix": "" }, { "first": "N", "middle": [], "last": "Burkhart", "suffix": "" }, { "first": "A", "middle": [], "last": "Minassians", "suffix": "" }, { "first": "G", "middle": [], "last": "Dervisoglu", "suffix": "" }, { "first": "L", "middle": [], "last": "Gutnik", "suffix": "" }, { "first": "M", "middle": [ "B" ], "last": "Haick", "suffix": "" }, { "first": "C", "middle": [], "last": "Ho", "suffix": "" }, { "first": "M", "middle": [], "last": "Koplow", "suffix": "" }, { "first": "J", "middle": [], "last": "Mangold", "suffix": "" }, { "first": "S", "middle": [], "last": "Robinson", "suffix": "" }, { "first": "M", "middle": [], "last": "Rosa", "suffix": "" }, { "first": "M", "middle": [], "last": "Schwartz", "suffix": "" }, { "first": "C", "middle": [], "last": "Sims", "suffix": "" }, { "first": "H", "middle": [], "last": "Stoffregen", "suffix": "" }, { "first": "A", "middle": [], "last": "Waterbury", "suffix": "" }, { "first": "E", "middle": [ "S" ], "last": "Leland", "suffix": "" }, { "first": "T", "middle": [], "last": "Pering", "suffix": "" }, { "first": "P", "middle": [ "K" ], "last": "Wright", "suffix": "" } ], "year": null, "venue": "Proc. 21st", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "C. R. Baker, K. Armijo, S. Belka, M. Benhabib, V. Bhargava, N. Burkhart, A. Der Minassians, G. Dervisoglu, L. Gutnik, M. B. Haick, C. Ho, M. Koplow, J. Mangold, S. Robinson, M. Rosa, M. Schwartz, C. Sims, H. Stoffregen, A. Waterbury, E. S. Leland, T. Pering, and P. K. Wright, \"Wireless sensor networks for home health care,\" in Proc. 21st", "links": null }, "BIBREF3": { "ref_id": "b3", "title": "Advanced Information Networking and Applications Workshops", "authors": [ { "first": "", "middle": [], "last": "Int", "suffix": "" }, { "first": "", "middle": [], "last": "Conf", "suffix": "" } ], "year": 2007, "venue": "", "volume": "", "issue": "", "pages": "832--837", "other_ids": {}, "num": null, "urls": [], "raw_text": "Int. Conf. Advanced Information Networking and Applications Workshops, Niagara Falls, Canada, 2007, May 21-23, pp. 832-837.", "links": null }, "BIBREF4": { "ref_id": "b4", "title": "Integration of wireless sensor network services into other home and industrial networks using device profile for web services (DPWS)", "authors": [ { "first": "A", "middle": [], "last": "Sleman", "suffix": "" }, { "first": "R", "middle": [], "last": "Moeller", "suffix": "" } ], "year": 2007, "venue": "Proc. 3rd Int. Conf. Information and Communication Technologies: From Theory to Applications", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "A. Sleman and R. Moeller, \"Integration of wireless sensor network services into other home and industrial networks using device profile for web services (DPWS),\" in Proc. 3rd Int. Conf. Information and Communication Technologies: From Theory to Applications, Damascus, Syria, 2008, Apr. 07-[12p. 1-5.", "links": null }, "BIBREF5": { "ref_id": "b5", "title": "Wireless sensor network based E-health system-Implementation and experimental results", "authors": [ { "first": "H", "middle": [], "last": "Yan", "suffix": "" }, { "first": "H", "middle": [], "last": "Huo", "suffix": "" }, { "first": "Y", "middle": [], "last": "Xu", "suffix": "" }, { "first": "M", "middle": [], "last": "Gidlund", "suffix": "" } ], "year": 2010, "venue": "IEEE Trans. Consumer Electronics", "volume": "56", "issue": "4", "pages": "2288--2295", "other_ids": {}, "num": null, "urls": [], "raw_text": "H. Yan, H. Huo, Y. Xu, and M. Gidlund, \"Wireless sensor network based E-health system-Implementation and experimental results,\" IEEE Trans. Consumer Electronics, vol. 56, no. 4, pp. 2288-2295, Nov. 2010.", "links": null }, "BIBREF6": { "ref_id": "b6", "title": "Robust environmental sound recognition for home automation", "authors": [ { "first": "J.-C", "middle": [], "last": "Wang", "suffix": "" }, { "first": "H.-P", "middle": [], "last": "Lee", "suffix": "" }, { "first": "J.-F", "middle": [], "last": "Wang", "suffix": "" }, { "first": "C.-B", "middle": [], "last": "Lin", "suffix": "" } ], "year": 2008, "venue": "IEEE Trans. Automation Science and Engineering", "volume": "5", "issue": "1", "pages": "25--31", "other_ids": {}, "num": null, "urls": [], "raw_text": "J.-C. Wang, H.-P. Lee, J.-F. Wang, and C.-B. Lin, \"Robust environmental sound recognition for home automation,\" IEEE Trans. Automation Science and Engineering, vol. 5, no. 1, pp. 25-31, Jan. 2008.", "links": null }, "BIBREF7": { "ref_id": "b7", "title": "Source extraction in bandwidth constrained wireless sensor networks", "authors": [ { "first": "H", "middle": [], "last": "Chen", "suffix": "" }, { "first": "C", "middle": [ "K" ], "last": "Tse", "suffix": "" }, { "first": "J", "middle": [], "last": "Feng", "suffix": "" } ], "year": 2008, "venue": "IEEE Trans. Circuits and Systems II: Express Briefs", "volume": "55", "issue": "9", "pages": "947--951", "other_ids": {}, "num": null, "urls": [], "raw_text": "H. Chen, C. K. Tse, and J. Feng, \"Source extraction in bandwidth constrained wireless sensor networks,\" IEEE Trans. Circuits and Systems II: Express Briefs, vol. 55, no. 9, pp. 947-951, Sep. 2008.", "links": null }, "BIBREF8": { "ref_id": "b8", "title": "Impact of topology on performance and energy efficiency in wireless sensor networks for source extraction", "authors": [ { "first": "H", "middle": [], "last": "Chen", "suffix": "" }, { "first": "C", "middle": [ "K" ], "last": "Tse", "suffix": "" }, { "first": "J", "middle": [], "last": "Feng", "suffix": "" } ], "year": 2009, "venue": "IEEE Trans. Parallel and Distributed Systems", "volume": "20", "issue": "6", "pages": "886--897", "other_ids": {}, "num": null, "urls": [], "raw_text": "H. Chen, C. K. Tse, and J. Feng, \"Impact of topology on performance and energy efficiency in wireless sensor networks for source extraction,\" IEEE Trans. Parallel and Distributed Systems, vol. 20, no. 6, pp. 886-897, Jun. 2009.", "links": null }, "BIBREF9": { "ref_id": "b9", "title": "Blind source extraction for a combined fixed and wireless sensor network", "authors": [ { "first": "B", "middle": [], "last": "Bloemendal", "suffix": "" }, { "first": "J", "middle": [ "V D" ], "last": "Laar", "suffix": "" }, { "first": "P", "middle": [], "last": "Sommen", "suffix": "" } ], "year": 2012, "venue": "Proc. 20th European Signal Processing Conference", "volume": "", "issue": "", "pages": "1264--1268", "other_ids": {}, "num": null, "urls": [], "raw_text": "B. Bloemendal, J. v. d. Laar, and P. Sommen, \"Blind source extraction for a combined fixed and wireless sensor network,\" in Proc. 20th European Signal Processing Conference, Bucharest, Romania, 2012, Aug. 27-31, pp. 1264-1268.", "links": null }, "BIBREF10": { "ref_id": "b10", "title": "Blind separation of underdetermined convolutive mixtures using their time-frequency Representation", "authors": [ { "first": "A", "middle": [], "last": "Aissa-El-Bey", "suffix": "" }, { "first": "K", "middle": [], "last": "Abed-Mraim", "suffix": "" }, { "first": "Y", "middle": [], "last": "Grenier", "suffix": "" } ], "year": 2007, "venue": "IEEE Trans. Audio, Speech, Lang. Process", "volume": "15", "issue": "", "pages": "1540--1550", "other_ids": {}, "num": null, "urls": [], "raw_text": "A. Aissa-El-Bey, K. Abed-Mraim, and Y. Grenier, \"Blind separation of underdetermined convolutive mixtures using their time-frequency Representation,\" IEEE Trans. Audio, Speech, Lang. Process., vol. 15, pp. 1540-1550, Jul. 2007.", "links": null }, "BIBREF11": { "ref_id": "b11", "title": "Underdetermined blind sparse source separation for arbitrarily arranged multiple sensors", "authors": [ { "first": "S", "middle": [], "last": "Araki", "suffix": "" }, { "first": "H", "middle": [], "last": "Sawada", "suffix": "" }, { "first": "R", "middle": [], "last": "Mukai", "suffix": "" }, { "first": "S", "middle": [], "last": "Makino", "suffix": "" } ], "year": 2007, "venue": "Signal Processing", "volume": "87", "issue": "", "pages": "1833--1847", "other_ids": {}, "num": null, "urls": [], "raw_text": "S. Araki, H. Sawada, R. Mukai, and S. Makino, \"Underdetermined blind sparse source separation for arbitrarily arranged multiple sensors,\" Signal Processing, vol. 87, pp. 1833-1847, Feb. 2007.", "links": null }, "BIBREF12": { "ref_id": "b12", "title": "MAP-based underdetermined blind source separation of convolutive mixtures by hierarchical clustering and l 1 -norm minimization", "authors": [ { "first": "S", "middle": [], "last": "Winter", "suffix": "" }, { "first": "W", "middle": [], "last": "Kellermann", "suffix": "" }, { "first": "H", "middle": [], "last": "Sawada", "suffix": "" }, { "first": "S", "middle": [], "last": "Makino", "suffix": "" } ], "year": 2007, "venue": "EURASIP Journal on Advances in Signal Processing", "volume": "12", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "S. Winter, W. Kellermann, H. Sawada, and S. Makino, \"MAP-based underdetermined blind source separation of convolutive mixtures by hierarchical clustering and l 1 -norm minimization,\" EURASIP Journal on Advances in Signal Processing, vol. 2007, Article ID 24717, 12 pages.", "links": null }, "BIBREF13": { "ref_id": "b13", "title": "Statistical Learning Theory", "authors": [ { "first": "V", "middle": [], "last": "Vapnik", "suffix": "" } ], "year": 1998, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "V. Vapnik, Statistical Learning Theory, New York: Wiley, 1998.", "links": null }, "BIBREF14": { "ref_id": "b14", "title": "Chip design of MFCC extraction for speech recognition", "authors": [ { "first": "J", "middle": [ "C" ], "last": "Wang", "suffix": "" }, { "first": "J", "middle": [ "F" ], "last": "Wang", "suffix": "" }, { "first": "Y", "middle": [ "S" ], "last": "Weng", "suffix": "" } ], "year": 2002, "venue": "Integration, the VLSI journal", "volume": "32", "issue": "", "pages": "111--131", "other_ids": {}, "num": null, "urls": [], "raw_text": "J. C. Wang, J. F. Wang, and Y. S. Weng, \"Chip design of MFCC extraction for speech recognition,\" Integration, the VLSI journal, vol. 32, pp. 111-131, 2002.", "links": null }, "BIBREF15": { "ref_id": "b15", "title": "Speaker verification using sequence discriminant support vector machines", "authors": [ { "first": "V", "middle": [], "last": "Wan", "suffix": "" }, { "first": "S", "middle": [], "last": "Renals", "suffix": "" } ], "year": 2005, "venue": "IEEE Trans. Speech and Audio Processing", "volume": "13", "issue": "", "pages": "203--210", "other_ids": {}, "num": null, "urls": [], "raw_text": "V. Wan and S. Renals, \"Speaker verification using sequence discriminant support vector machines,\" IEEE Trans. Speech and Audio Processing, vol. 13, pp.203-210, Mar. 2005.", "links": null }, "BIBREF16": { "ref_id": "b16", "title": "Performance measurement in blind audio source separation", "authors": [ { "first": "E", "middle": [], "last": "Vincent", "suffix": "" }, { "first": "R", "middle": [], "last": "Gribonval", "suffix": "" }, { "first": "C", "middle": [], "last": "Fevotte", "suffix": "" } ], "year": 2006, "venue": "IEEE Trans. Audio, Speech Lang. Process", "volume": "14", "issue": "", "pages": "1462--1469", "other_ids": {}, "num": null, "urls": [], "raw_text": "E. Vincent, R. Gribonval and C. Fevotte, \"Performance measurement in blind audio source separation,\" IEEE Trans. Audio, Speech Lang. Process., vol. 14, pp. 1462-1469, 2006.", "links": null }, "BIBREF17": { "ref_id": "b17", "title": "Strategies for automatic segmentation of audio data", "authors": [ { "first": "T", "middle": [], "last": "Kemp", "suffix": "" }, { "first": "M", "middle": [], "last": "Schmidt", "suffix": "" }, { "first": "M", "middle": [], "last": "Westphal", "suffix": "" }, { "first": "A", "middle": [], "last": "Waibel", "suffix": "" } ], "year": 2000, "venue": "Proc. Int. Conf. Acoust., Speech, Signal Process", "volume": "3", "issue": "", "pages": "1423--1426", "other_ids": {}, "num": null, "urls": [], "raw_text": "T. Kemp, M. Schmidt, M. Westphal, and A. Waibel, \"Strategies for automatic segmentation of audio data,\" in Proc. Int. Conf. Acoust., Speech, Signal Process., vol. 3, 2000, pp. 1423-1426.", "links": null } }, "ref_entries": { "FIGREF0": { "num": null, "type_str": "figure", "text": "\u56e0\u70ba\u672c\u8ad6\u6587\u8981\u5c0d\u7531 Level-Ratio \u4ee5\u53ca Phase-Difference \u6240\u7d44\u6210\u7684\u6a23\u672c\u4f5c\u7fa4\u805a\u5206\u5272\u3002\u6240 \u4ee5\uff0c\u5f97\u51fa\u4e86\u6df7\u5408\u8a0a\u865f\u6a23\u672c\u5728\u6975\u5ea6\u7a00\u758f\u7684\u60c5\u5f62\u4e0b\u8868\u73fe\u7684\u578b\u5f0f\u5f8c\uff0c\u6211\u5011\u5c07\u5f0f\u5b50(13)\u4ee3\u5165\u5f0f\u5b50 (6)\u548c\u5f0f\u5b50(7)\uff0c\u770b\u770b\u82e5\u5229\u7528\u9019\u7a2e\u5f62\u614b\u7684\u6a23\u672c\u53bb\u5b9a\u7fa9 Level-Ratio \u548c Phase-Difference \u9019\u5169\u7a2e", "uris": null }, "TABREF3": { "type_str": "table", "html": null, "text": "Separated Sound 17.5794 db 17.7916 db 18.7986 db \u8868\u4e00\u5217\u51fa\u4e86\u4e09\u7a2e\u4e0d\u540c\u7684 DOA \u5dee\u7570\u503c(40\u00b0, 80\u00b0, \u548c 160\u00b0)\u7684\u5e73\u5747 SIR\u3002\u5be6\u9a57\u7d50\u679c\u986f \u793a\u8d8a\u5927\u7684 DOA \u5dee\u7570\u901a\u5e38\u5c0e\u81f4\u8d8a\u9ad8\u7684 SIR\uff0c\u9032\u800c\u4fdd\u8b49\u4e86\u8f03\u4f73\u7684\u5206\u96e2\u8a0a\u865f\u3002\u9019\u6a23\u7684\u6e2c\u8a66\u652f \u6301\u4e86\u7cfb\u7d71\u9078\u64c7\u5177\u6709\u6700\u5927 DOA \u503c\u7684\u611f\u6e2c\u7bc0\u9ede\u4f86\u9032\u884c\u8a0a\u865f\u5206\u96e2\u3002 \u5728\u9a57\u8b49\u7cfb\u7d71\u65b9\u9762\uff0c\u6211\u5011\u6bd4\u8f03\u4e86\u6df7\u5408\u8a0a\u865f\u548c\u5206\u96e2\u8a0a\u865f\u7684\u9a57\u8b49\u8868\u73fe\u3002\u5728\u5206\u96e2\u8a0a\u865f\u65b9\u9762\uff0c \u7cfb\u7d71\u6703\u6aa2\u8996\u5728\u5169\u500b\u8a0a\u865f\u4e2d\uff0c\u662f\u5426\u6709\u5f9e\u5c6c\u65bc\u76ee\u6a19\u985e\u5225\u7684\u5206\u96e2\u7a0b\u5e8f\u6240\u7522\u751f\u7684\u8a0a\u865f\u3002\u82e5\u9019\u5169\u500b \u8a0a\u865f\u7686\u4e0d\u70ba\u76ee\u6a19\u8072\u97f3\uff0c\u5247\u5c07\u6b64\u8072\u97f3\u8a0a\u865f\u6b78\u5c6c\u5728\u975e\u76ee\u6a19\u985e\u5225\u3002\u5c0d\u6bcf\u500b\u97f3\u6a94\uff0c\u6211\u5011\u5f9e\u6bcf\u500b\u8072 \u97f3\u8a0a\u865f\u7684\u97f3\u6846\u53d6\u51fa 13 \u7dad MFCCs\uff0c\u4ee5\u53ca\u6240\u6709\u97f3\u6846\u7684 MFCCs \u7684\u5e73\u5747\u503c\u53ca\u6a19\u6e96\u5dee\u4f86\u4f5c\u70ba\u8072 \u97f3\u7279\u5fb5\u3002\u540c\u6642\uff0c\u6211\u5011\u5f9e\u6bcf\u500b\u8a0a\u865f\u7684 3 \u968e\u5c0f\u6ce2\u5305\u5206\u89e3\u6a39 (Three-Level Wavelet Packet Decomposition Tree) \u4e2d\u53d6\u51fa 16 \u500b\u8cbb\u820d\u723e\u5206\u6578\uff0c\u5f97\u5230\u7e3d\u6578\u70ba 42 \u7dad\u7684\u7279\u5fb5\u5411\u91cf\u3002\u6211\u5011\u4f7f\u7528 F-Score [17]\u91cf\u6e2c\u65b9\u6cd5\u4f86\u8a55\u6e2c\u6211\u5011\u7684\u7cfb\u7d71\u3002 \u8868\u4e8c\u3001\u57fa\u6e96\u7cfb\u7d71\u548c\u63d0\u51fa\u7684\u7cfb\u7d71\u5728\u9a57\u8b49\u8868\u73fe\u7684\u6bd4\u8f03", "num": null, "content": "
\u8868\u4e00\u3001\u4e0d\u540c DOA \u5dee\u7570\u503c\u7684\u5e73\u5747 SIR
DOA Difference40\u00b080\u00b0160\u00b0
14] \uff0c \u6211 \u5011 \u5c07 \u8072 \u97f3 \u8a0a \u865f \u5207 \u6210 \u77ed \u6642 \u7a97 (Short-Time Window)\u4e26\u5c0d\u6bcf\u500b\u77ed\u6642\u7a97\u9032\u884c\u5feb\u901f\u5085\u7acb\u8449\u8f49\u63db(Fast Fourier Transform) \u3002 \u7136\u5f8c\u5c07\u983b\u8b5c\u6240\u5305\u542b\u7684\u80fd\u91cf\u6620\u5c04\u5230\u4f7f\u7528\u4e09\u89d2\u6ffe\u6ce2\u5668\u983b\u5e36\u7684\u6885\u723e\u5c3a\u5ea6\u4e0a\u3002\u5728\u6bcf\u500b\u6885\u723e\u6ffe\u6ce2\u983b \u5e36\u4e0a\uff0c\u6211\u5011\u8a08\u7b97\u5c0d\u6578\u80fd\u91cf\uff0c\u6700\u5f8c\u7528\u96e2\u6563\u9918\u5f26\u8f49\u63db(Discrete Cosine Transform)\u4f86\u5f97\u5230 MFCCs\u3002 2\u3001\u8cbb\u820d\u723e\u5206\u6578\u7a7a\u9593 \u8cbb\u820d\u723e\u5206\u6578\u4f7f\u7528\u4e00\u7d44\u53c3\u6578\u751f\u6210\u6a21\u578b\uff0c\u6b64\u6a21\u578b\u5c07\u4e00\u4e32\u5e8f\u5217\u6620\u5c04\u5230\u56fa\u5b9a\u7dad\u5ea6\u7a7a\u9593\u7684\u55ae\u4e00 \u9ede\u4e0a\uff0c\u4f8b\u5982\uff1a\u5206\u6578\u7a7a\u9593 [15]\u3002\u85c9\u7531\u751f\u6210\u6a21\u578b\u7684\u4f3c\u7136\u503c(Likelihood)\u5206\u6578\u4f86\u57f7\u884c\u6b64\u6620\u5c04\u3002 \u5728\u9019\u9805\u7814\u7a76\u4e2d\uff0c\u6211\u5011\u63a8\u5c0e\u51fa\u8cbb\u820d\u723e\u5206\u6578\u7528\u4f86\u5c07\u7bc0\u9ede\u7684\u6a5f\u7387\u5206\u4f48\u5c0d\u6620\u5230\u8072\u97f3\u7684\u5c0f\u6ce2\u5305\u5206 \u89e3\u3002\u6839\u64da\u7d93\u9a57\uff0c\u6bcf\u500b\u7bc0\u9ede\u88ab\u8996\u4f5c\u4e00\u7a2e\u55ae\u4e00\u9ad8\u65af\u5206\u4f48\u3002 \u7d66\u5b9a\u4e00\u500b\u53c3\u6578\u96c6\u70ba\u03b8\u7684\u751f\u6210\u6a21\u578b p(X|\u03b8)\uff0c\u6211\u5011\u53ef\u4ee5\u85c9\u7531\u4e0b\u5217\u7684\u5f0f\u5b50\u5f97\u5230\u76f8\u95dc\u7684\u5206\u6578 \u7a7a\u9593\u3002 F-score SIR of Sound Class Separated Sounds Separated Sounds Mixed Sounds (Baseline) (Proposed System)
Doorbell Ringing0.000.460.92
Glass BreakingF \u03a8,f) 0.82 ( X=F(f(p(| 0.80 \u03b8 X)))0.92(22)
Door Knocking0.230.000.32
\u5728\u4e0a\u5f0f\u4e2d\uff0c\u03a8\u8868\u793a\u5206\u6578\u5411\u91cf\uff1bf(p(X|\u03b8))\u8868\u793a\u5206\u6578\u53c3\u6578\uff0c\u6b64\u5206\u6578\u53c3\u6578\u70ba\u672c\u7bc7\u8ad6\u6587\u4e2d\u751f
\u6210\u6a21\u578b\u7684\u5c0d\u6578\u4f3c\u7136\u51fd\u6578\uff1bF\u8868\u793a\u5c07\u5206\u6578\u53c3\u6578\u5c0d\u61c9\u5230\u5206\u6578\u7a7a\u9593\u7684\u5206\u6578\u904b\u7b97\u3002 \u8868\u4e8c\u70ba\u4e09\u500b\u76ee\u6a19\u8072\u97f3\u985e\u5225\u7684\u6df7\u5408\u8072\u97f3\u53ca\u5206\u96e2\u8072\u97f3\u7684\u9a57\u8b49\u7d50\u679c\u6bd4\u8f03\u3002\u6211\u5011\u53ef\u4ee5\u770b\u51fa\u5728
\u9a57\u8b49\u7cfb\u7d71\u4e2d\uff0c\u7528\u5206\u96e2\u8a0a\u865f\u6bd4\u7528\u6df7\u5408\u8a0a\u865f\u7684\u6548\u679c\u4f86\u7684\u597d\u5f88\u591a\u3002\u6b64\u5916\uff0c\u6211\u5011\u63d0\u51fa\u7684 CBSS \u7cfb \u4e94\u3001\u5be6\u9a57\u7d50\u679c \u7d71\u8868\u73fe\u4e5f\u6bd4\u57fa\u6e96\u7cfb\u7d71\u512a\u7570\u3002
\u6211\u5011\u5be6\u9a57\u7684\u7b2c\u4e00\u500b\u76ee\u6a19\u5728\u8b49\u660e\u6211\u5011\u7684\u8072\u97f3\u5206\u96e2\u968e\u6bb5\u7684\u8868\u73fe\u3002\u63a5\u8457\uff0c\u6211\u5011\u85c9\u7531\u6bd4\u8f03\u6df7 \u5408\u548c\u5206\u96e2\u7684\u8072\u97f3\u7684\u9a57\u8b49\u7d50\u679c\u4f86\u5c55\u793a\u8072\u97f3\u5206\u96e2\u5c0d\u8072\u97f3\u9a57\u8b49\u7684\u8ca2\u737b\u3002\u5728\u6b64\u7814\u7a76\u4e2d\uff0c\u6211\u5011\u5229\u7528 \u516d\u3001\u7d50\u8ad6
\u4e09\u7a2e\u6211\u5011\u611f\u8208\u8da3\u7684\u8072\u97f3\u8a0a\u865f\u7a2e\u985e\u4f86\u4f5c\u70ba\u76ee\u6a19\u8072\u97f3\uff0c\u4f8b\u5982\uff1a\u9580\u9234\u97ff\u3001\u73bb\u7483\u7834\u88c2\u548c\u6572\u9580\u8072\u3002 \u5728\u672c\u7bc7\u8ad6\u6587\u4e2d\uff0c\u6211\u5011\u63cf\u8ff0\u4e86\u4e00\u500b\u5728\u5bb6\u5ead\u81ea\u52d5\u5316\u4e2d\uff0c\u57fa\u65bc\u7121\u7dda\u611f\u6e2c\u7db2\u8def\u4e4b\u6df7\u5408\u8072\u97f3\u4e8b
\u6211\u5011\u9084\u5b9a\u7fa9\u4e86\u56db\u500b\u4e0d\u5e0c\u671b\u5f97\u5230\u7684\u8072\u97f3\uff0c\u5305\u542b\u8c93\u53eb\u3001\u72d7\u5420\u3001\u5f48\u92fc\u7434\u53ca\u4eba\u8aaa\u8a71\u4f86\u7576\u4f5c\u975e\u76ee\u6a19 \u4ef6\u5206\u96e2\u548c\u9a57\u8b49\u7cfb\u7d71\u3002\u6211\u5011\u8aaa\u660e\u4e86\u65cb\u7a4d\u76f2\u8a0a\u865f\u6e90\u5206\u96e2\u53ef\u4ee5\u88ab\u7528\u4f86\u5206\u96e2\u6df7\u548c\u8072\u97f3\u4e8b\u4ef6\u8a0a\u865f\u3002
\u8072\u97f3\u3002\u5728\u8a13\u7df4\u968e\u6bb5\uff0c\u6211\u5011\u7528\u5f9e\u9019\u4e9b\u985e\u5225\u7684\u4e7e\u6de8\u8072\u97f3\u53d6\u51fa\u7684\u7279\u5fb5\u503c\u4f86\u8a13\u7df4 SVM \u5206\u985e\u5668\u3002 \u9664\u4e86\u6df7\u5408\u8072\u97f3\u5206\u96e2\uff0c\u6211\u5011\u63a1\u7528\u652f\u63f4\u5411\u91cf\u6a5f\u4f86\u9032\u884c\u8072\u97f3\u9a57\u8b49\u3002\u6240\u4f7f\u7528\u7684\u7279\u5fb5\u96c6\u5305\u542b MFCCs
\u6b64\u5916\uff0c20 \u500b\u5f9e\u76ee\u6a19\u985e\u5225\u9078\u51fa\u7684\u4e7e\u6de8\u8072\u97f3\u6a94\u6848\u53ca 30 \u500b\u5f9e\u975e\u76ee\u6a19\u985e\u5225\u9078\u51fa\u7684\u8072\u97f3\u7247\u6bb5\u4e5f\u88ab \u548c Fisher Scores\u3002\u6211\u5011\u7684\u5be6\u9a57\u986f\u793a\uff0c\u6240\u63d0\u51fa\u7684\u6df7\u5408\u8072\u97f3\u5206\u96e2\u67b6\u69cb\u986f\u8457\u5730\u589e\u9032\u4e86\u8072\u97f3\u9a57\u8b49
\u7528\u4f86\u8a13\u7df4 SVM \u5206\u985e\u5668\u3002\u6211\u5011\u7684\u7cfb\u7d71\u4e2d\u4f7f\u7528\u7684\u8072\u97f3\u9577\u5ea6\u662f 1 \u79d2\uff0c\u53d6\u6a23\u7387\u70ba 8 kHz\u3002 \u7684\u7d50\u679c\u3002
\u6211\u5011\u7528\u76ee\u6a19\u8072\u97f3\u548c\u4eba\u8aaa\u8a71\u8072\u7684\u6df7\u5408\u8a0a\u865f\u4f86\u6e2c\u8a66\u7cfb\u7d71\u3002\u5728\u7b2c\u4e00\u500b\u5be6\u9a57\u4e2d\uff0c\u6211\u5011\u7528\u4e0d\u540c \u7684 DOA \u5dee\u7570\u503c\u4f86\u8a55\u4f30\u8072\u6e90\u6578\u91cf\u4f30\u8a08\u548c\u5206\u96e2\u7684\u8868\u73fe\u3002\u6211\u5011\u63a1\u7528\u8a0a\u865f\u5e72\u64fe\u6bd4(Signal-To-\u53c3\u8003\u6587\u737b
Interference Ratio, SIR)[16] \u4f86\u8a55\u50f9\u5206\u96e2\u7684\u8868\u73fe\u3002
" } } } }