seg_detector_representer.py

import cv2
import numpy as np
from shapely.geometry import Polygon
import pyclipper
from concern.config import Configurable, State

class SegDetectorRepresenter(Configurable):
    thresh = State(default=0.3)
    box_thresh = State(default=0.7)
    max_candidates = State(default=100)
    dest = State(default='binary')

    def __init__(self, cmd={}, **kwargs):
        self.load_all(**kwargs)
        self.min_size = 3
        self.scale_ratio = 0.4
        if 'debug' in cmd:
            self.debug = cmd['debug']
        if 'thresh' in cmd:
            self.thresh = cmd['thresh']
        if 'box_thresh' in cmd:
            self.box_thresh = cmd['box_thresh']
        if 'dest' in cmd:
            self.dest = cmd['dest']

    def represent(self, batch, _pred, is_output_polygon=False):
        '''
        batch: (image, polygons, ignore_tags
        batch: a dict produced by dataloaders.
            image: tensor of shape (N, C, H, W).
            polygons: tensor of shape (N, K, 4, 2), the polygons of objective regions.
            ignore_tags: tensor of shape (N, K), indicates whether a region is ignorable or not.
            shape: the original shape of images.
            filename: the original filenames of images.
        pred:
            binary: text region segmentation map, with shape (N, 1, H, W)
            thresh: [if exists] thresh hold prediction with shape (N, 1, H, W)
            thresh_binary: [if exists] binarized with threshhold, (N, 1, H, W)
        '''
        images = batch['image']
        if isinstance(_pred, dict):
            pred = _pred[self.dest]
        else:
            pred = _pred
        segmentation = self.binarize(pred)
        boxes_batch = []
        scores_batch = []
        for batch_index in range(images.size(0)):
            height, width = batch['shape'][batch_index]
            if is_output_polygon:
                boxes, scores = self.polygons_from_bitmap(
                    pred[batch_index],
                    segmentation[batch_index], width, height)
            else:
                boxes, scores = self.boxes_from_bitmap(
                    pred[batch_index],
                    segmentation[batch_index], width, height)
            boxes_batch.append(boxes)
            scores_batch.append(scores)
        return boxes_batch, scores_batch
    
    def binarize(self, pred):
        return pred > self.thresh

    def polygons_from_bitmap(self, pred, _bitmap, dest_width, dest_height):
        '''
        _bitmap: single map with shape (1, H, W),
            whose values are binarized as {0, 1}
        '''

        assert _bitmap.size(0) == 1
        bitmap = _bitmap.cpu().numpy()[0]  # The first channel
        pred = pred.cpu().detach().numpy()[0]
        height, width = bitmap.shape
        boxes = []
        scores = []

        contours, _ = cv2.findContours(
            (bitmap*255).astype(np.uint8),
            cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)

        for contour in contours[:self.max_candidates]:
            epsilon = 0.01 * cv2.arcLength(contour, True)
            approx = cv2.approxPolyDP(contour, epsilon, True)
            points = approx.reshape((-1, 2))
            if points.shape[0] < 4:
                continue
            # _, sside = self.get_mini_boxes(contour)
            # if sside < self.min_size:
            #     continue
            score = self.box_score_fast(pred, points.reshape(-1, 2))
            if self.box_thresh > score:
                continue
            
            if points.shape[0] > 2:
                box = self.unclip(points, unclip_ratio=2.0)
                if len(box) > 1:
                    continue
            else:
                continue
            box = box.reshape(-1, 2)
            _, sside = self.get_mini_boxes(box.reshape((-1, 1, 2)))
            if sside < self.min_size + 2:
                continue

            if not isinstance(dest_width, int):
                dest_width = dest_width.item()
                dest_height = dest_height.item()
            
            box[:, 0] = np.clip(
                np.round(box[:, 0] / width * dest_width), 0, dest_width)
            box[:, 1] = np.clip(
                np.round(box[:, 1] / height * dest_height), 0, dest_height)
            boxes.append(box.tolist())
            scores.append(score)
        return boxes, scores

    def boxes_from_bitmap(self, pred, _bitmap, dest_width, dest_height):
        '''
        _bitmap: single map with shape (1, H, W),
            whose values are binarized as {0, 1}
        '''
        
        assert _bitmap.size(0) == 1
        bitmap = _bitmap.cpu().numpy()[0]  # The first channel
        pred = pred.cpu().detach().numpy()[0]
        height, width = bitmap.shape
        contours, _ = cv2.findContours(
            (bitmap*255).astype(np.uint8),
            cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
        num_contours = min(len(contours), self.max_candidates)
        boxes = np.zeros((num_contours, 4, 2), dtype=np.int16)
        scores = np.zeros((num_contours,), dtype=np.float32)

        for index in range(num_contours):
            contour = contours[index]
            points, sside = self.get_mini_boxes(contour)
            if sside < self.min_size:
                continue
            points = np.array(points)
            score = self.box_score_fast(pred, points.reshape(-1, 2))
            if self.box_thresh > score:
                continue
        
            box = self.unclip(points).reshape(-1, 1, 2)
            box, sside = self.get_mini_boxes(box)
            if sside < self.min_size + 2:
                continue
            box = np.array(box)
            if not isinstance(dest_width, int):
                dest_width = dest_width.item()
                dest_height = dest_height.item()
            
            box[:, 0] = np.clip(
                np.round(box[:, 0] / width * dest_width), 0, dest_width)
            box[:, 1] = np.clip(
                np.round(box[:, 1] / height * dest_height), 0, dest_height)
            boxes[index, :, :] = box.astype(np.int16)
            scores[index] = score
        return boxes, scores

    def unclip(self, box, unclip_ratio=1.5):
        poly = Polygon(box)
        distance = poly.area * unclip_ratio / poly.length
        offset = pyclipper.PyclipperOffset()
        offset.AddPath(box, pyclipper.JT_ROUND, pyclipper.ET_CLOSEDPOLYGON)
        expanded = np.array(offset.Execute(distance))
        return expanded

    def get_mini_boxes(self, contour):
        bounding_box = cv2.minAreaRect(contour)
        points = sorted(list(cv2.boxPoints(bounding_box)), key=lambda x: x[0])

        index_1, index_2, index_3, index_4 = 0, 1, 2, 3
        if points[1][1] > points[0][1]:
            index_1 = 0
            index_4 = 1
        else:
            index_1 = 1
            index_4 = 0
        if points[3][1] > points[2][1]:
            index_2 = 2
            index_3 = 3
        else:
            index_2 = 3
            index_3 = 2

        box = [points[index_1], points[index_2],
               points[index_3], points[index_4]]
        return box, min(bounding_box[1])

    def box_score_fast(self, bitmap, _box):
        h, w = bitmap.shape[:2]
        box = _box.copy()
        xmin = np.clip(np.floor(box[:, 0].min()).astype(np.int), 0, w - 1)
        xmax = np.clip(np.ceil(box[:, 0].max()).astype(np.int), 0, w - 1)
        ymin = np.clip(np.floor(box[:, 1].min()).astype(np.int), 0, h - 1)
        ymax = np.clip(np.ceil(box[:, 1].max()).astype(np.int), 0, h - 1)

        mask = np.zeros((ymax - ymin + 1, xmax - xmin + 1), dtype=np.uint8)
        box[:, 0] = box[:, 0] - xmin
        box[:, 1] = box[:, 1] - ymin
        cv2.fillPoly(mask, box.reshape(1, -1, 2).astype(np.int32), 1)
        return cv2.mean(bitmap[ymin:ymax+1, xmin:xmax+1], mask)[0]