""" This file contains functions that are used to perform data augmentation. """ from turtle import reset import cv2 import io import torch import numpy as np import scipy.misc from PIL import Image from rembg.bg import remove from torchvision.models import detection from lib.pymaf.core import constants from lib.pymaf.utils.streamer import aug_matrix from lib.common.cloth_extraction import load_segmentation from torchvision import transforms def load_img(img_file): img = cv2.imread(img_file, cv2.IMREAD_UNCHANGED) if len(img.shape) == 2: img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR) if not img_file.endswith("png"): img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) else: img = cv2.cvtColor(img, cv2.COLOR_RGBA2BGR) return img def get_bbox(img, det): input = np.float32(img) input = (input / 255.0 - (0.5, 0.5, 0.5)) / (0.5, 0.5, 0.5) # TO [-1.0, 1.0] input = input.transpose(2, 0, 1) # TO [3 x H x W] bboxes, probs = det(torch.from_numpy(input).float().unsqueeze(0)) probs = probs.unsqueeze(3) bboxes = (bboxes * probs).sum(dim=1, keepdim=True) / probs.sum( dim=1, keepdim=True) bbox = bboxes[0, 0, 0].cpu().numpy() return bbox def get_transformer(input_res): image_to_tensor = transforms.Compose([ transforms.Resize(input_res), transforms.ToTensor(), transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)) ]) mask_to_tensor = transforms.Compose([ transforms.Resize(input_res), transforms.ToTensor(), transforms.Normalize((0.0, ), (1.0, )) ]) image_to_pymaf_tensor = transforms.Compose([ transforms.Resize(size=224), transforms.Normalize(mean=constants.IMG_NORM_MEAN, std=constants.IMG_NORM_STD) ]) image_to_pixie_tensor = transforms.Compose([transforms.Resize(224)]) def image_to_hybrik_tensor(img): # mean img[0].add_(-0.406) img[1].add_(-0.457) img[2].add_(-0.480) # std img[0].div_(0.225) img[1].div_(0.224) img[2].div_(0.229) return img return [ image_to_tensor, mask_to_tensor, image_to_pymaf_tensor, image_to_pixie_tensor, image_to_hybrik_tensor ] def process_image(img_file, hps_type, input_res=512, device=None, seg_path=None): """Read image, do preprocessing and possibly crop it according to the bounding box. If there are bounding box annotations, use them to crop the image. If no bounding box is specified but openpose detections are available, use them to get the bounding box. """ [ image_to_tensor, mask_to_tensor, image_to_pymaf_tensor, image_to_pixie_tensor, image_to_hybrik_tensor ] = get_transformer(input_res) img_ori = load_img(img_file) in_height, in_width, _ = img_ori.shape M = aug_matrix(in_width, in_height, input_res * 2, input_res * 2) # from rectangle to square img_for_crop = cv2.warpAffine(img_ori, M[0:2, :], (input_res * 2, input_res * 2), flags=cv2.INTER_CUBIC) # detection for bbox detector = detection.maskrcnn_resnet50_fpn(pretrained=True) detector.eval() predictions = detector( [torch.from_numpy(img_for_crop).permute(2, 0, 1) / 255.])[0] human_ids = torch.logical_and( predictions["labels"] == 1, predictions["scores"] == predictions["scores"].max()).nonzero().squeeze(1) bbox = predictions["boxes"][human_ids, :].flatten().detach().cpu().numpy() width = bbox[2] - bbox[0] height = bbox[3] - bbox[1] center = np.array([(bbox[0] + bbox[2]) / 2.0, (bbox[1] + bbox[3]) / 2.0]) scale = max(height, width) / 180 if hps_type == 'hybrik': img_np = crop_for_hybrik(img_for_crop, center, np.array([scale * 180, scale * 180])) else: img_np, cropping_parameters = crop(img_for_crop, center, scale, (input_res, input_res)) with torch.no_grad(): buf = io.BytesIO() Image.fromarray(img_np).save(buf, format='png') img_pil = Image.open(io.BytesIO(remove( buf.getvalue()))).convert("RGBA") # for icon img_rgb = image_to_tensor(img_pil.convert("RGB")) img_mask = torch.tensor(1.0) - (mask_to_tensor(img_pil.split()[-1]) < torch.tensor(0.5)).float() img_tensor = img_rgb * img_mask # for hps img_hps = img_np.astype(np.float32) / 255. img_hps = torch.from_numpy(img_hps).permute(2, 0, 1) if hps_type == 'bev': img_hps = img_np[:, :, [2, 1, 0]] elif hps_type == 'hybrik': img_hps = image_to_hybrik_tensor(img_hps).unsqueeze(0).to(device) elif hps_type != 'pixie': img_hps = image_to_pymaf_tensor(img_hps).unsqueeze(0).to(device) else: img_hps = image_to_pixie_tensor(img_hps).unsqueeze(0).to(device) # uncrop params uncrop_param = { 'center': center, 'scale': scale, 'ori_shape': img_ori.shape, 'box_shape': img_np.shape, 'crop_shape': img_for_crop.shape, 'M': M } if not (seg_path is None): segmentations = load_segmentation(seg_path, (in_height, in_width)) seg_coord_normalized = [] for seg in segmentations: coord_normalized = [] for xy in seg['coordinates']: xy_h = np.vstack((xy[:, 0], xy[:, 1], np.ones(len(xy)))).T warped_indeces = M[0:2, :] @ xy_h[:, :, None] warped_indeces = np.array(warped_indeces).astype(int) warped_indeces.resize((warped_indeces.shape[:2])) # cropped_indeces = crop_segmentation(warped_indeces, center, scale, (input_res, input_res), img_np.shape) cropped_indeces = crop_segmentation(warped_indeces, (input_res, input_res), cropping_parameters) indices = np.vstack( (cropped_indeces[:, 0], cropped_indeces[:, 1])).T # Convert to NDC coordinates seg_cropped_normalized = 2 * (indices / input_res) - 1 # Don't know why we need to divide by 50 but it works ¯\_(ツ)_/¯ (probably some scaling factor somewhere) # Divide only by 45 on the horizontal axis to take the curve of the human body into account seg_cropped_normalized[:, 0] = (1 / 40) * seg_cropped_normalized[:, 0] seg_cropped_normalized[:, 1] = (1 / 50) * seg_cropped_normalized[:, 1] coord_normalized.append(seg_cropped_normalized) seg['coord_normalized'] = coord_normalized seg_coord_normalized.append(seg) return img_tensor, img_hps, img_ori, img_mask, uncrop_param, seg_coord_normalized return img_tensor, img_hps, img_ori, img_mask, uncrop_param def get_transform(center, scale, res): """Generate transformation matrix.""" h = 200 * scale t = np.zeros((3, 3)) t[0, 0] = float(res[1]) / h t[1, 1] = float(res[0]) / h t[0, 2] = res[1] * (-float(center[0]) / h + .5) t[1, 2] = res[0] * (-float(center[1]) / h + .5) t[2, 2] = 1 return t def transform(pt, center, scale, res, invert=0): """Transform pixel location to different reference.""" t = get_transform(center, scale, res) if invert: t = np.linalg.inv(t) new_pt = np.array([pt[0] - 1, pt[1] - 1, 1.]).T new_pt = np.dot(t, new_pt) return np.around(new_pt[:2]).astype(np.int16) def crop(img, center, scale, res): """Crop image according to the supplied bounding box.""" # Upper left point ul = np.array(transform([0, 0], center, scale, res, invert=1)) # Bottom right point br = np.array(transform(res, center, scale, res, invert=1)) new_shape = [br[1] - ul[1], br[0] - ul[0]] if len(img.shape) > 2: new_shape += [img.shape[2]] new_img = np.zeros(new_shape) # Range to fill new array new_x = max(0, -ul[0]), min(br[0], len(img[0])) - ul[0] new_y = max(0, -ul[1]), min(br[1], len(img)) - ul[1] # Range to sample from original image old_x = max(0, ul[0]), min(len(img[0]), br[0]) old_y = max(0, ul[1]), min(len(img), br[1]) new_img[new_y[0]:new_y[1], new_x[0]:new_x[1]] = img[old_y[0]:old_y[1], old_x[0]:old_x[1]] if len(img.shape) == 2: new_img = np.array(Image.fromarray(new_img).resize(res)) else: new_img = np.array( Image.fromarray(new_img.astype(np.uint8)).resize(res)) return new_img, (old_x, new_x, old_y, new_y, new_shape) def crop_segmentation(org_coord, res, cropping_parameters): old_x, new_x, old_y, new_y, new_shape = cropping_parameters new_coord = np.zeros((org_coord.shape)) new_coord[:, 0] = new_x[0] + (org_coord[:, 0] - old_x[0]) new_coord[:, 1] = new_y[0] + (org_coord[:, 1] - old_y[0]) new_coord[:, 0] = res[0] * (new_coord[:, 0] / new_shape[1]) new_coord[:, 1] = res[1] * (new_coord[:, 1] / new_shape[0]) return new_coord def crop_for_hybrik(img, center, scale): inp_h, inp_w = (256, 256) trans = get_affine_transform(center, scale, 0, [inp_w, inp_h]) new_img = cv2.warpAffine(img, trans, (int(inp_w), int(inp_h)), flags=cv2.INTER_LINEAR) return new_img def get_affine_transform(center, scale, rot, output_size, shift=np.array([0, 0], dtype=np.float32), inv=0): def get_dir(src_point, rot_rad): """Rotate the point by `rot_rad` degree.""" sn, cs = np.sin(rot_rad), np.cos(rot_rad) src_result = [0, 0] src_result[0] = src_point[0] * cs - src_point[1] * sn src_result[1] = src_point[0] * sn + src_point[1] * cs return src_result def get_3rd_point(a, b): """Return vector c that perpendicular to (a - b).""" direct = a - b return b + np.array([-direct[1], direct[0]], dtype=np.float32) if not isinstance(scale, np.ndarray) and not isinstance(scale, list): scale = np.array([scale, scale]) scale_tmp = scale src_w = scale_tmp[0] dst_w = output_size[0] dst_h = output_size[1] rot_rad = np.pi * rot / 180 src_dir = get_dir([0, src_w * -0.5], rot_rad) dst_dir = np.array([0, dst_w * -0.5], np.float32) src = np.zeros((3, 2), dtype=np.float32) dst = np.zeros((3, 2), dtype=np.float32) src[0, :] = center + scale_tmp * shift src[1, :] = center + src_dir + scale_tmp * shift dst[0, :] = [dst_w * 0.5, dst_h * 0.5] dst[1, :] = np.array([dst_w * 0.5, dst_h * 0.5]) + dst_dir src[2:, :] = get_3rd_point(src[0, :], src[1, :]) dst[2:, :] = get_3rd_point(dst[0, :], dst[1, :]) if inv: trans = cv2.getAffineTransform(np.float32(dst), np.float32(src)) else: trans = cv2.getAffineTransform(np.float32(src), np.float32(dst)) return trans def corner_align(ul, br): if ul[1] - ul[0] != br[1] - br[0]: ul[1] = ul[0] + br[1] - br[0] return ul, br def uncrop(img, center, scale, orig_shape): """'Undo' the image cropping/resizing. This function is used when evaluating mask/part segmentation. """ res = img.shape[:2] # Upper left point ul = np.array(transform([0, 0], center, scale, res, invert=1)) # Bottom right point br = np.array(transform(res, center, scale, res, invert=1)) # quick fix ul, br = corner_align(ul, br) # size of cropped image crop_shape = [br[1] - ul[1], br[0] - ul[0]] new_img = np.zeros(orig_shape, dtype=np.uint8) # Range to fill new array new_x = max(0, -ul[0]), min(br[0], orig_shape[1]) - ul[0] new_y = max(0, -ul[1]), min(br[1], orig_shape[0]) - ul[1] # Range to sample from original image old_x = max(0, ul[0]), min(orig_shape[1], br[0]) old_y = max(0, ul[1]), min(orig_shape[0], br[1]) img = np.array(Image.fromarray(img.astype(np.uint8)).resize(crop_shape)) new_img[old_y[0]:old_y[1], old_x[0]:old_x[1]] = img[new_y[0]:new_y[1], new_x[0]:new_x[1]] return new_img def rot_aa(aa, rot): """Rotate axis angle parameters.""" # pose parameters R = np.array([[np.cos(np.deg2rad(-rot)), -np.sin(np.deg2rad(-rot)), 0], [np.sin(np.deg2rad(-rot)), np.cos(np.deg2rad(-rot)), 0], [0, 0, 1]]) # find the rotation of the body in camera frame per_rdg, _ = cv2.Rodrigues(aa) # apply the global rotation to the global orientation resrot, _ = cv2.Rodrigues(np.dot(R, per_rdg)) aa = (resrot.T)[0] return aa def flip_img(img): """Flip rgb images or masks. channels come last, e.g. (256,256,3). """ img = np.fliplr(img) return img def flip_kp(kp, is_smpl=False): """Flip keypoints.""" if len(kp) == 24: if is_smpl: flipped_parts = constants.SMPL_JOINTS_FLIP_PERM else: flipped_parts = constants.J24_FLIP_PERM elif len(kp) == 49: if is_smpl: flipped_parts = constants.SMPL_J49_FLIP_PERM else: flipped_parts = constants.J49_FLIP_PERM kp = kp[flipped_parts] kp[:, 0] = -kp[:, 0] return kp def flip_pose(pose): """Flip pose. The flipping is based on SMPL parameters. """ flipped_parts = constants.SMPL_POSE_FLIP_PERM pose = pose[flipped_parts] # we also negate the second and the third dimension of the axis-angle pose[1::3] = -pose[1::3] pose[2::3] = -pose[2::3] return pose def normalize_2d_kp(kp_2d, crop_size=224, inv=False): # Normalize keypoints between -1, 1 if not inv: ratio = 1.0 / crop_size kp_2d = 2.0 * kp_2d * ratio - 1.0 else: ratio = 1.0 / crop_size kp_2d = (kp_2d + 1.0) / (2 * ratio) return kp_2d def generate_heatmap(joints, heatmap_size, sigma=1, joints_vis=None): ''' param joints: [num_joints, 3] param joints_vis: [num_joints, 3] return: target, target_weight(1: visible, 0: invisible) ''' num_joints = joints.shape[0] device = joints.device cur_device = torch.device(device.type, device.index) if not hasattr(heatmap_size, '__len__'): # width height heatmap_size = [heatmap_size, heatmap_size] assert len(heatmap_size) == 2 target_weight = np.ones((num_joints, 1), dtype=np.float32) if joints_vis is not None: target_weight[:, 0] = joints_vis[:, 0] target = torch.zeros((num_joints, heatmap_size[1], heatmap_size[0]), dtype=torch.float32, device=cur_device) tmp_size = sigma * 3 for joint_id in range(num_joints): mu_x = int(joints[joint_id][0] * heatmap_size[0] + 0.5) mu_y = int(joints[joint_id][1] * heatmap_size[1] + 0.5) # Check that any part of the gaussian is in-bounds ul = [int(mu_x - tmp_size), int(mu_y - tmp_size)] br = [int(mu_x + tmp_size + 1), int(mu_y + tmp_size + 1)] if ul[0] >= heatmap_size[0] or ul[1] >= heatmap_size[1] \ or br[0] < 0 or br[1] < 0: # If not, just return the image as is target_weight[joint_id] = 0 continue # # Generate gaussian size = 2 * tmp_size + 1 # x = np.arange(0, size, 1, np.float32) # y = x[:, np.newaxis] # x0 = y0 = size // 2 # # The gaussian is not normalized, we want the center value to equal 1 # g = np.exp(- ((x - x0) ** 2 + (y - y0) ** 2) / (2 * sigma ** 2)) # g = torch.from_numpy(g.astype(np.float32)) x = torch.arange(0, size, dtype=torch.float32, device=cur_device) y = x.unsqueeze(-1) x0 = y0 = size // 2 # The gaussian is not normalized, we want the center value to equal 1 g = torch.exp(-((x - x0)**2 + (y - y0)**2) / (2 * sigma**2)) # Usable gaussian range g_x = max(0, -ul[0]), min(br[0], heatmap_size[0]) - ul[0] g_y = max(0, -ul[1]), min(br[1], heatmap_size[1]) - ul[1] # Image range img_x = max(0, ul[0]), min(br[0], heatmap_size[0]) img_y = max(0, ul[1]), min(br[1], heatmap_size[1]) v = target_weight[joint_id] if v > 0.5: target[joint_id][img_y[0]:img_y[1], img_x[0]:img_x[1]] = \ g[g_y[0]:g_y[1], g_x[0]:g_x[1]] return target, target_weight