mirrors
/
yolov7
mirror of https://github.com/WongKinYiu/yolov7.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Fix image_data_creator.py : truncated objects in lower part of image: fix effective height computation 

datasets.py : Remove empty annotated image entries
weights file moved to HDD
pull/2071/head
hanoch 2025-02-27 12:45:47 +02:00
parent 52320239d4
commit 8766685f94
2 changed files with 73 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
train.py
View File

@ -987,6 +987,9 @@ tir_od_overfit.yaml
# 3 classe renew yolov7999 list
--workers 8 --device 0 --batch-size 24 --data data/tir_od_center_roi_aug_list_train_cls.yaml --img 640 640 --weights /mnt/Data/hanoch/tir_frames_rois/yolov7.pt --cfg cfg/training/yolov7.yaml --name yolov7 --hyp hyp.tir_od.tiny_aug_gamma_scaling_before_mosaic_rnd_scaling_no_ota.yaml --adam --norm-type single_image_percentile_0_1 --input-channels 1 --linear-lr --epochs 100 --gamma-aug-prob 0.1 --cache-images --image-weights
--workers 8 --device 0 --batch-size 24 --data data/tir_od_center_roi_aug_list_train_cls.yaml --img 640 640 --weights /mnt/Data/hanoch/tir_frames_rois/yolov7.pt --cfg cfg/training/yolov7.yaml --name yolov7 --hyp hyp.tir_od.tiny_aug_gamma_scaling_before_mosaic_rnd_scaling_no_ota.yaml --adam --norm-type single_image_percentile_0_1 --input-channels 1 --linear-lr --epochs 100 --gamma-aug-prob 0.1 --cache-images --image-weights --loss-weight
#########################################################
Extended model for higher resolution YOLO7E6
# --workers 8 --device 0 --batch-size 8 --data data/tir_od_center_roi_aug_list_full_res.yaml --weights /mnt/Data/hanoch/tir_frames_rois/yolov7-e6.pt --img-size [768, 1024] --cfg cfg/deploy/yolov7-e6.yaml --name yolov7e --hyp hyp.tir_od.aug_gamma_scaling_before_mosaic_rnd_scaling_e6_full_res.yaml --adam --norm-type single_image_percentile_0_1 --input-channels 1 --linear-lr --epochs 2 --gamma-aug-prob 0.3 --cache-images --rect

90
utils/datasets.py
View File

@ -39,6 +39,36 @@ eps = 1e-5
import pandas as pd
def flatten(lst): return [x for l in lst for x in l]
def clip_boxes_to_border(boxes_array, border):
"""
Clips bounding boxes to stay within the given border.
Args:
boxes_array (np.ndarray): A NumPy array of shape (N, 5),
where N is the number of boxes.
Format: [class_id, x_min, y_min, x_max, y_max].
border (tuple): Border box as (x_min, y_min, x_max, y_max).
Returns:
np.ndarray: Clipped bounding boxes with the same shape (N, 5).
"""
if boxes_array.shape[1] != 5:
raise ValueError("The array must have shape (N, 5) with [class_id, x_min, y_min, x_max, y_max]")
x_min_border, y_min_border, x_max_border, y_max_border = border
# Copy the original array to avoid modifying input
clipped_boxes = np.copy(boxes_array)
# Clip coordinates while keeping class_id unchanged
clipped_boxes[:, 1] = np.clip(clipped_boxes[:, 1], x_min_border, x_max_border) # x_min
clipped_boxes[:, 2] = np.clip(clipped_boxes[:, 2], y_min_border, y_max_border) # y_min
clipped_boxes[:, 3] = np.clip(clipped_boxes[:, 3], x_min_border, x_max_border) # x_max
clipped_boxes[:, 4] = np.clip(clipped_boxes[:, 4], y_min_border, y_max_border) # y_max
return clipped_boxes
help_url = 'https://github.com/ultralytics/yolov5/wiki/Train-Custom-Data'
img_formats = ['bmp', 'jpg', 'jpeg', 'png', 'tif', 'tiff', 'dng', 'webp', 'mpo'] # acceptable image suffixes
vid_formats = ['mov', 'avi', 'mp4', 'mpg', 'mpeg', 'm4v', 'wmv', 'mkv'] # acceptable video suffixes
@ -644,24 +674,36 @@ class LoadImagesAndLabels(Dataset): # for training/testing
cache.pop('hash') # remove hash
cache.pop('version') # remove version
labels, shapes, self.segments = zip(*cache.values())
shapes = list(shapes)
# #@@HK TODO adding truncation ratio increase here
# if labels.shape[1] > 5:
# labels = labels[:,:5]
# self.truncation_ratio = labels[:,5]
self.labels = list(labels)
self.shapes = np.array(shapes, dtype=np.float64)
self.img_files = list(cache.keys()) # update
self.label_files = img2label_paths(cache.keys()) # update
# [l for ix,l in enumerate(self.labels) if (l[:, 1:] >1).all()]
if single_cls:
for x in self.labels:
x[:, 0] = 0
if nm > 0:
print('Remove missing annotations file avoiding unlabeled images that would considered as BG')
for ix in range(len(self.labels) - 1, -1, -1): # safe remove by reverrse iteration #enumerate(self.labels):
if (self.labels[ix][:, 1:] > 1).any() or self.labels[ix].size < 5:
del self.labels[ix]
del self.img_files[ix]
del self.label_files[ix]
del shapes[ix]
self.shapes = np.array(shapes, dtype=np.float64)
n = len(shapes) # number of images
bi = np.floor(np.arange(n) / batch_size).astype(int) # batch index
nb = bi[-1] + 1 # number of batches
self.batch = bi # batch index of image
self.n = n
self.indices = range(n)
self.mosiac_no = 0
# Rectangular Training
if self.rect:
@ -806,7 +848,7 @@ class LoadImagesAndLabels(Dataset): # for training/testing
# return self
def __getitem__(self, index):
index = self.indices[index] # linear, shuffled, or image_weights
index = self.indices[index] # linear, shuffled, or image_weights HK@@ since the fi
file_type = os.path.basename(self.img_files[index]).split('.')[-1].lower()
@ -1437,14 +1479,18 @@ def hist_equalize(img, clahe=True, bgr=False):
yuv[:, :, 0] = cv2.equalizeHist(yuv[:, :, 0]) # equalize Y channel histogram
return cv2.cvtColor(yuv, cv2.COLOR_YUV2BGR if bgr else cv2.COLOR_YUV2RGB) # convert YUV image to RGB
def load_mosaic(self, index, filling_value, file_type='tiff'):
# loads images in a 4-mosaic
self.mosiac_no += 1
labels4, segments4 = [], []
s = self.img_size
yc, xc = [int(random.uniform(-x, 2 * s + x)) for x in self.mosaic_border] # mosaic center x, y
indices = [index] + random.choices(self.indices, k=3) # 3 additional image indices
# debug = True
# if debug:
# indices[1] = 17106 #TIR2_V60_Jan21_Test51D_ML_RD_IL_26_12_2021_15_50_27_FS_210_XGA_0001_0100_ROTEM_left_roi_210_85
for i, index in enumerate(indices):
# Load image
img, _, (h, w) = load_image(self, index)
@ -1456,8 +1502,10 @@ def load_mosaic(self, index, filling_value, file_type='tiff'):
percentile=self.percentile, beta=self.beta)
# import tifffile
# tifffile.imwrite(os.path.join('/home/hanoch/projects/tir_od/output',
# 'img_projective_' + str(self.img_files[index].split('/')[-1].split('.tiff')[0]) +'.tiff'), img)
# tifffile.imwrite(os.path.join('/mnt/Data/hanoch/runs/output',
# str(self.mosiac_no) + 'img_mosiac_' + '_crop_no_' + str(i) + '_'\
# +str(self.img_files[index].split('/')[-1].split('.tiff')[0]) +'.tiff'),
# img)
# place img in img4
if i == 0: # top left
@ -1468,8 +1516,8 @@ def load_mosaic(self, index, filling_value, file_type='tiff'):
x1a, y1a, x2a, y2a = max(xc - w, 0), max(yc - h, 0), xc, yc # xmin, ymin, xmax, ymax (large image)
x1b, y1b, x2b, y2b = w - (x2a - x1a), h - (y2a - y1a), w, h # xmin, ymin, xmax, ymax (small image)
elif i == 1: # top right
x1a, y1a, x2a, y2a = xc, max(yc - h, 0), min(xc + w, s * 2), yc
x1b, y1b, x2b, y2b = 0, h - (y2a - y1a), min(w, x2a - x1a), h
x1a, y1a, x2a, y2a = xc, max(yc - h, 0), min(xc + w, s * 2), yc # dest image
x1b, y1b, x2b, y2b = 0, h - (y2a - y1a), min(w, x2a - x1a), h # src image
elif i == 2: # bottom left
x1a, y1a, x2a, y2a = max(xc - w, 0), yc, xc, min(s * 2, yc + h)
x1b, y1b, x2b, y2b = w - (x2a - x1a), 0, w, min(y2a - y1a, h)
@ -1486,6 +1534,7 @@ def load_mosaic(self, index, filling_value, file_type='tiff'):
if labels.size:
labels[:, 1:] = xywhn2xyxy(labels[:, 1:], w, h, padw, padh) # normalized xywh to pixel xyxy format
segments = [xyn2xy(x, w, h, padw, padh) for x in segments]
labels = clip_boxes_to_border(labels, (x1a, y1a, x2a, y2a))
labels4.append(labels)
segments4.extend(segments)
@ -1511,18 +1560,19 @@ def load_mosaic(self, index, filling_value, file_type='tiff'):
# border to remove
# import tifffile
# tifffile.imwrite(os.path.join('/home/hanoch/projects/tir_od/output',
# 'img_projective_' + str(self.img_files[indices[0]].split('/')[-1].split('.tiff')[0]) +'.tiff'), img)
#
# tifffile.imwrite(os.path.join('/home/hanoch/projects/tir_od/output',
# 'img_projective_' + str(
# self.img_files[indices[1]].split('/')[-1].split('.tiff')[0]) + '.tiff'), img)
# tifffile.imwrite(os.path.join('/home/hanoch/projects/tir_od/output',
# 'img_projective_' + str(
# self.img_files[indices[2]].split('/')[-1].split('.tiff')[0]) + '.tiff'), img)
# tifffile.imwrite(os.path.join('/home/hanoch/projects/tir_od/output',
# 'img_projective_' + str(
# self.img_files[indices[3]].split('/')[-1].split('.tiff')[0]) + '.tiff'), img)
# tifffile.imwrite(os.path.join('/mnt/Data/hanoch/runs/output',
# str(self.mosiac_no) + '_img_mosaic' + '_' + str(
# self.img_files[indices[0]].split('/')[-1].split('.tiff')[0]) +'.tiff'),
# img4)
# tifffile.imwrite(os.path.join('/home/hanoch/projects/tir_od/output',
# 'img_mosiac_' + str(self.mosiac_no) + '_' + str(
# self.img_files[indices[1]].split('/')[-1].split('.tiff')[0]) + '.tiff'), img)
# tifffile.imwrite(os.path.join('/home/hanoch/projects/tir_od/output',
# 'img_mosiac_' + str(self.mosiac_no) + '_' + str(
# self.img_files[indices[2]].split('/')[-1].split('.tiff')[0]) + '.tiff'), img)
# tifffile.imwrite(os.path.join('/home/hanoch/projects/tir_od/output',
# 'img_mosiac_' + str(self.mosiac_no) + '_' + str(
# self.img_files[indices[3]].split('/')[-1].split('.tiff')[0]) + '.tiff'), img)
return img4, labels4