File size: 2,224 Bytes
9d283bc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
#!/bin/bash

# Get script dir
SCRIPT_DIR=$(cd $(dirname $0); pwd)
LABELME_DIR=$SCRIPT_DIR/../export_annotated
COCO_DIR=$SCRIPT_DIR/../export_coco

# Create necessary directories
mkdir -p $COCO_DIR/train
mkdir -p $COCO_DIR/val
mkdir -p $COCO_DIR/annotations

# Clean train, val, and annotations directories by removing all files
rm -rf $COCO_DIR/train/*
rm -rf $COCO_DIR/val/*
rm -rf $COCO_DIR/annotations/*

# Determine the number of files for validation (20%)
LABELME_FILES=$(find "$LABELME_DIR" -name "*.json")
NUM_FILES=$(echo "$LABELME_FILES" | wc -l)
VAL_COUNT=$(echo "$NUM_FILES * 0.2" | bc | awk '{print int($1+0.5)}')

# Shuffle files and split into train and val sets
SHUFFLED_FILES=$(echo "$LABELME_FILES" | shuf)
VAL_FILES=$(echo "$SHUFFLED_FILES" | head -n $VAL_COUNT)
TRAIN_FILES=$(echo "$SHUFFLED_FILES" | tail -n +$(($VAL_COUNT + 1)))

# Copy validation files and their corresponding images to val
echo "$VAL_FILES" | while read -r val_file; do
  cp "$val_file" "$COCO_DIR/val/"
  
  # Copy the corresponding image file (assuming it has the same base name)
  image_file="${val_file%.json}.png"  # Change the extension to match your images
  if [[ -f "$image_file" ]]; then
    cp "$image_file" "$COCO_DIR/val/"
  fi
done

# Copy training files and their corresponding images to train
echo "$TRAIN_FILES" | while read -r train_file; do
  cp "$train_file" "$COCO_DIR/train/"
  
  # Copy the corresponding image file (assuming it has the same base name)
  image_file="${train_file%.json}.png"  # Change the extension to match your images
  if [[ -f "$image_file" ]]; then
    cp "$image_file" "$COCO_DIR/train/"
  fi
done

# Convert Labelme JSON to COCO format for train and val
labelme2coco $COCO_DIR/train $COCO_DIR/annotations
mv $COCO_DIR/annotations/dataset.json $COCO_DIR/annotations/train.json
labelme2coco $COCO_DIR/val $COCO_DIR/annotations
mv $COCO_DIR/annotations/dataset.json $COCO_DIR/annotations/val.json

# Remove the Labelme JSON files from train and val directories after conversion
find $COCO_DIR/train -name "*.json" -type f -delete
find $COCO_DIR/val -name "*.json" -type f -delete

echo "Conversion to COCO format completed, and JSON files have been deleted from train and val directories."