libccv Examples

Index

1. Introduction
2. List of examples
3. Expected output
4. Dependencies

Introduction

This application runs several examples using the libccv library. Some of them compare the obtained results vs expected result, giving an assertion if the values are different. In others, the user needs to check the results saved in the SD card.

The contents of the testFiles folder (data and samples) must be copied to the SD card (/mnt/sdcard/).

List of examples

1. Canny edge detector.
   1. Files required:
      • samples/blackbox.png
      • data/blackbox.canny.bin
   2. Output:
      • File: samples/test1.png (1 if border has been detected, 0 if not)
2. Pedestrian cascade detector using Integral Channel Features.
   1. Files required:
      • samples/pedestrian.icf
      • samples/pedestrian2.png
   2. Output:
      • Text: Number of pedestrians detected. For each detection, classification confidence and x, y, width and height of the rectangle.
3. Matrix addition operation.
4. Sobel operations.
   1. Files required:
      • samples/chessbox.bmp
      • data/chessbox.sobel.x.bin
      • data/chessbox.sobel.y.bin
      • data/chessbox.sobel.u.bin
      • data/chessbox.sobel.v.bin
      • data/chessbox.sobel.x.3.bin
      • data/chessbox.sobel.y.3.bin
      • data/chessbox.sobel.x.5.bin
   2. Output:
      • samples/test4.png (partial derivative on y within 5x5 window)
5. Otsu threshold calculation.
6. Image contrast modification.
   1. Files required:
      • samples/nature.bmp
      • data/nature.contrast.0.5.bin
      • data/nature.contrast.1.5.bin
   2. Output:
      • samples/test6_1.png (reducing contrast to 0.5)
      • samples/test6_2.png (increasing contrast 1.5 times)
7. Perspective transform on a picture.
   1. Files required:
      • sampls/chessbox.bmp
      • data/chessbox.perspective.transform.bin
   2. Output:
      • samples/test7.png (picture rotated along y-axis for 30º)
8. Histogram of gradients calculation.
   1. Files required:
      • samples/nature.bmp
   2. Output:
      • samples/test8.png (values obtained)
9. Scale Invariant Feature Transform.
   1. Files required:
      • samples/book.png
   2. Output:
      • Text: List of keypoints matching.
10. CBLAS matrix multiplication.
11. Convolutional network of 11x11 on 225x225 with uniform weights.
12. Convolutional network of 5x5 on 27x27 with non-uniform weights.
13. Convolutional network of 5x5x4 on 27x27x8 partitioned by 2.
14. Stroke Width Transform (text detection).
    1. Files required:
       • samples/text-detect.png
    2. Output:
       • Text: Number of texts detected. For each detection, x, y, width and height of the rectangle.
15. Brightness Binary Feature (face cascade detection).
    1. Files required:
       • samples/face (folder with cascade.txt and stage-0.txt to stage-15.txt files)
       • samples/suit.png
    2. Output:
       • Text: Number of faces detected. For each detection, classification confidence and x, y, width and height of the rectangle.

Expected output

UART: Thread 1 created
UART: Starting run
UART: Example 1, canny and write png
UART: Value of image: rows - 500, cols - 500, type - 1074794497
UART: Example 2, pedestrian cascade detector (several days)
UART: Loading files
UART: Detecting pedestrians
UART: Results:
UART: 20 36 50 152 0.074491
UART: total : 1 detected
UART: Example 3, matrix addition
UART: Example 4, sobel
UART: Example 5, otsu threshold
UART: Example 6, image contrast
UART: Example 7, perspective transform
UART: Example 8, HOG
UART: Example 9, SIFT
UART: Pictures read
UART: Keypoints calculated obj 1179
UART: Keypoints calculated img 1179
UART: 73.833488 2.090770 => 73.833488 2.090770
UART: 30.298519 2.555899 => 30.298519 2.555899
UART: 10.025833 2.887055 => 10.025833 2.887055
…………
…………
…………
UART: 105.916695 222.729172 => 105.916695 222.729172
UART: 105.916695 222.729172 => 105.916695 222.729172
UART: 117.846596 238.603943 => 117.846596 238.603943
UART: 75.254646 144.538391 => 75.254646 144.538391
UART: 338x289 on 338x289
UART: 1179 keypoints out of 1179 are matched
UART: elpased time : 0
UART: Example 10, CBLAS mat multiplication
UART: Example 11, Convolutional network
UART: Example 12, convolutional network of 5x5 on 27x27 with non-uniform weigh
UART: ts
UART: Example 13, convolutional network of 5x5x4 on 27x27x8 partitioned by 2
UART: Example 14, Text detect
UART: 64 179 104 13
UART: 182 178 67 13
UART: 89 197 136 18
UART: 183 197 34 13
UART: total : 4 detected
UART: Example 15, Detect faces
UART: 144 59 53 53 2.364821
UART: total : 1 detected
UART: Tests ended

Dependencies

• Crypto
• SDCardIO
• libccv