Execute the following steps for this recipe:
- First, download the Caltech101 images from http://www.vision.caltech.edu/Image_Datasets/Caltech101/ and decompress it. There are image objects corresponding to 101 classes, although we have used a subset of them, corresponding to only 12 class labels (namely brain, butterfly, Buddha, chair, elephant, laptop, piano, pigeon, pizza, pyramid, rhino, sunflower) to speed up the preprocessing and training phases, but, you can train using all the classes if you want.
- Read the images corresponding to the 10 different classes (they get extracted in separate folders) from the respective folders and extract the HOG images/descriptors from the images using the following code. Print the number ...
