Bulk extract Caffe features (to be continued) (Python, C + +, Matlab)

This article refers to the following:

Instant Recognition with Caffe
Extracting Features

Caffe Python feature Extraction

Caffe Practice 4--Use Python to bulk extract Caffe Compute features--by banana melody
Caffe Exercise 3 Use the C + + function provided by Caffe to extract image features in batches--by banana melody

Caffe Python batch extraction of image features
Caffe Python Batch extraction image feature-sequel
Caffe C + + extract picture features

Shicai C + + Caffe extraction Features

Caffe Source Code Modification: Extract any picture features

Matlab batch extraction of CNN Features

The framework of this article on how to extract features in batches is as follows:
1. Preparation of data and corresponding preparation work
2. Initializing the network
3. Read the image list
4. Extract image features and save them in a specific format

Python method One
There are three main functions:
Initialize () Initialization of the network related
ReadList () Read the Extract image list
Extractfeatre () Extracts the characteristics of an image, saving it as a specified format

Where transformer needs to be set according to its own needs.

#encoding: Utf-8 #详情请查看http://www.cnblogs.com/louyihang-loves-baiyan/p/5078746.html import numpy as NP import Matplotlib.pyplot as PLT import OS import Caffe import sys import pickle import struct import sys,cv2 = '. /' # Run the model's prototxt deployprototxt = '/home/bids/caffe/caffe-master/changmiao/model/deploy.prototxt ' # corresponding loaded modelfile m
Odelfile = '/home/bids/caffe/caffe-master/changmiao/model/bvlc_reference_caffenet.caffemodel ' # meanfile can also be generated by itself Meanfile = ' python/caffe/imagenet/ilsvrc_2012_mean.npy ' # A list of images to extract imagelistfile = '/home/bids/caffe/caffe-master/　Changmiao/data/temp.txt ' Imagebasepath = '/home/bids/caffe/caffe-master/changmiao/data/cat ' #gpuID = 4 # Depending on your computer's GPU situation, postfix = '. Classify_allcar1716_fc6 ' # Initialization functions related to Operations Def initilize (): print ' Initilize ... ' Sys.path.insert (0, caffe_root + ' python ') Caffe.set_mode_gpu () # caffe.set_device (gpuid) net = Caffe. Net (Deployprototxt, Modelfile,caffe. TEST) return NET # extracts features and saves them as appropriate file Def extractfeatUre (imageList, net): # Adjust input data Accordingly, such as channel, dimension, etc. transformer = Caffe.io.Transformer ({' Data ': net.blobs[' data '].data.shap e}) transformer.set_transpose (' Data ', (2,0,1)) Transformer.set_mean (' Data ', np.load (Caffe_root + meanfile). Mean (1)  
    . Mean (1)) # mean pixel transformer.set_raw_scale (' data ', 255) transformer.set_channel_swap (' Data ', (2,1,0)) # Set NET to batch size of 1 if you have more pictures, set the appropriate batchsize net.blobs[' data '].reshape (1,3,227,227) #这里根据需要设定, if the network is inconsistent, you need to To adjust the num=0 #imageList = Os.listdir (Imagebasepath) for imagefile in imagelist:imagefile_abs = OS.PATH.J Oin (Imagebasepath, imagefile) print imagefile_abs net.blobs[' data '].data[...] = transformer.preprocess (' da Ta ', Caffe.io.load_image (imagefile_abs)) out = Net.forward () Fea_file = Imagefile_abs.replace ('. jpg ', POSTF
            (ix) NUM +=1 print ' num ', num, ' extract feature ', Fea_file with open (Fea_file, ' WB ') as F: For x in xrange (0, Net.blobs[' Fc6 '].data.shape[0]): For y in xrange (0, net.blobs[' fc6 '].data.shape[1]): F. Write (Struct.pack (' F ', net.blobs[' Fc6 '].data[x,y]) # Read file list def readimagelist (imagelistfile): ImageList = [] Wit H-Open (Imagelistfile, ' R ') as Fi:while (True): line = Fi.readline (). Strip (). Split () # every line is a IM Age file name if not Line:break imagelist.append (line[0]) print ' Read Imagel ist done Image num ', Len (imageList) return imageList if __name__ = = "__main__": Net = initilize () imageList = Readimagelist (imagelistfile) extractfeature (imageList, net)