faster rcnn 代码解析9

Pascal voc.py

#pascal_voc这个类主要用来组织输入的图片数据，存储图片的相关信息，但并不存储图片；而实际上，pascal_voc类是imdb类的一个子类
# --------------------------------------------------------
# Fast R-CNN
# Copyright (c) 2015 Microsoft
# Licensed under The MIT License [see LICENSE for details]
# Written by Ross Girshick
# --------------------------------------------------------
#该类继承于IMDB，用于负责数据交互部分
import os
from datasets.imdb import imdb
import datasets.ds_utils as ds_utils
import xml.etree.ElementTree as ET #ElementTree表示整个XML树，Element表示树上的单个节点
import numpy as np
import scipy.sparse
import scipy.io as sio
import utils.cython_bbox
import cPickle
import subprocess
import uuid
from voc_eval import voc_eval
from fast_rcnn.config import cfg

class pascal_voc(imdb):
    def __init__(self, image_set, year, devkit_path=None):
        #这个类用来组织输入的图片数据，但并没有将真实的图片存进去
        # 传进来的第一个参数为数据集名称（train，val，test），第
        # 二个参数为版本，如2007，
        # devkit_path暂时为空
        imdb.__init__(self, 'voc_' + year + '_' + image_set)
        #调用IMDB的构造函数，传入参数格式为‘voc_year_imageset'--例如voc_2007_train,
        # 其实就是记录了一下self._name,其余的为默认
        #其余默认参数有(self._num_classes,self._classes,self._image_index,self._obj_proposer,
        # self._roidb_handler,self.config)
        self._year = year
        self._image_set = image_set
        self._devkit_path = self._get_default_path() if devkit_path is None \
                            else devkit_path
        #devkit_path在不设定为none，此时self._devkit_path为py_faster_rcnn/data/ VOCdevkit+self._year
        self._data_path = os.path.join(self._devkit_path, 'VOC' + self._year)
        # self._data_path 为data / VOCdevkit +'year'/'VOC' + self._year
        self._classes = ('__background__','ant', 'butterfly','cicadas','dragonfly','ladybug','mantis','honeybee','fly','grasshopper','cricket')
        #self._class_to_ind里存的是{'__background__'：0,'aeroplane'：1.....}
        self._class_to_ind = dict(zip(self.classes, xrange(self.num_classes))) #给每一个类别分别赋予一个整数
        self._image_ext = '.jpg' #图片的扩展名
        self._image_index = self._load_image_set_index() #把所有的图片名称加载，放在list中，便于索引读取图片
        # Default to roidb handler
        # 得到ROI图片信息，重载IMDB中
        self._roidb_handler = self.selective_search_roidb
        self._salt = str(uuid.uuid4()) # UUID是128位的全局唯一标识符，通常由32字节的字符串表示。它可以保证时间和空间的唯一性，也称为GUID（C#）。它通过MAC地址、时间戳、命名空间、随机数、伪随机数来保证生成ID的唯一性。uuid4()——基于随机数；由伪随机数得到，有一定的重复概率，该概率可以计算出来。
        self._comp_id = 'comp4'

        # PASCAL specific config options PASCAL特殊配置选项
        self.config = {'cleanup'     : True,
                       'use_salt'    : True,
                       'use_diff'    : False,
                       'matlab_eval' : False,
                       'rpn_file'    : None,
                       'min_size'    : 2}

        assert os.path.exists(self._devkit_path), \
                'VOCdevkit path does not exist: {}'.format(self._devkit_path)
        assert os.path.exists(self._data_path), \
                'Path does not exist: {}'.format(self._data_path)

    def image_path_at(self, i):#重载了imdb.py中定义，返回图片所在全路径
        """
        Return the absolute path to image i in the image sequence.
        """
        return self.image_path_from_index(self._image_index[i])

    def image_path_from_index(self, index):
        """
        Construct an image path from the image's "index" identifier标识符.
        """
        image_path = os.path.join(self._data_path, 'JPEGImages',
                                  index + self._image_ext)#例如/home/ubuntu/py-faster-rcnn/data/VOCdevkit2007/VOC2007/JPEGImages/0.jpg
        assert os.path.exists(image_path), \
                'Path does not exist: {}'.format(image_path)
        return image_path

    def _load_image_set_index(self):
        """
        Load the indexes listed in this dataset's image set file.
        """
    #获取图片索引
        # Example path to image set file:
        # self._devkit_path + /VOCdevkit2007/VOC2007/ImageSets/Main/val.txt
        image_set_file = os.path.join(self._data_path, 'ImageSets', 'Main',
                                      self._image_set + '.txt')
        assert os.path.exists(image_set_file), \
                'Path does not exist: {}'.format(image_set_file)
        with open(image_set_file) as f:
            image_index = [x.strip() for x in f.readlines()]#移除字符串头尾指定的字符（默认为空格或换行符）或字符序列
        return image_index
    #返回的image_index为一个列表，包含该数据集图片名称信息(之前做VOC数据集时候就有在对应txt中，是没有.jpg后缀的，这是为了让你方便修改代码，制作自己的数据集)

    def _get_default_path(self):
        """
        Return the default path where PASCAL VOC is expected to be installed.返回Pascalvoc的路径
        """
        return os.path.join(cfg.DATA_DIR, 'VOCdevkit' + self._year)

    def gt_roidb(self): #获取ground_truth的roidb格式
        """
        Return the database of ground-truth regions of interest.得到ROI组成database

        This function loads/saves from/to a cache file to speed up future calls.
        """
        cache_file = os.path.join(self.cache_path, self.name + '_gt_roidb.pkl') #保存缓存文件，目录data/cache/voc_2007_train_gt_roidb.pkl
        if os.path.exists(cache_file):
            with open(cache_file, 'rb') as fid:#打开本地文件，并指定以二进制格式打开一个文件用于只读。文件指针将会放在文件的开头。这是默认模式。一般用于非文本文件如图片等。
                roidb = cPickle.load(fid) #加载fid到roidb中
            print '{} gt roidb loaded from {}'.format(self.name, cache_file)
            return roidb

        gt_roidb = [self._load_pascal_annotation(index) #使用_load_pascal_annotation()从XML中解析gt_roidb数据
                    for index in self.image_index]
        with open(cache_file, 'wb') as fid: #打开本地文件，并指定以二进制格式打开一个文件只用于写入。如果该文件已存在则将其覆盖。如果该文件不存在，创建新文件。一般用于非文本文件如图片等。
            cPickle.dump(gt_roidb, fid, cPickle.HIGHEST_PROTOCOL) #将roidb数据序列化保存到cache_file中（将python对象序列化保存到本地的文件）。一种高效的加载方式cPickle.HIGHEST_PROTOCOL，可使得节省80%空间。dump函数需要指定两个参数，第一个是需要序列化的python对象名称，第二个是本地的文件。cPickle.HIGHEST_PROTOCOL：An integer, the highest protocol version available. This value can be passed as a protocol value to functions dump() and dumps() as well as the Picklerconstructor.也可以使用-1表示。
        print 'wrote gt roidb to {}'.format(cache_file)

        return gt_roidb

    def selective_search_roidb(self):
        """
        Return the database of selective search regions of interest.返回候选ROI数据集
        Ground-truth ROIs are also included.

        This function loads/saves from/to a cache file to speed up future calls.
    没有RPN的fast-rcnn提取候选框的方式。返回的是提取出来的ROI以及图片的gt。
        这个函数在Faster-RCNN里面用不到，在fast-rcnn里面才会用到
        """
        cache_file = os.path.join(self.cache_path,
                                  self.name + '_selective_search_roidb.pkl')

        if os.path.exists(cache_file):#判断这个.pkl是否存在，如果存在调用该文件内容
            with open(cache_file, 'rb') as fid:
                roidb = cPickle.load(fid)#将selective_search_roidb.pkl加载到roidb中
            print '{} ss roidb loaded from {}'.format(self.name, cache_file)
            return roidb

        if int(self._year) == 2007 or self._image_set != 'test':
            gt_roidb = self.gt_roidb()
            ss_roidb = self._load_selective_search_roidb(gt_roidb)
            roidb = imdb.merge_roidbs(gt_roidb, ss_roidb)
        else:
            roidb = self._load_selective_search_roidb(None)
        with open(cache_file, 'wb') as fid:
            cPickle.dump(roidb, fid, cPickle.HIGHEST_PROTOCOL)
        print 'wrote ss roidb to {}'.format(cache_file)

        return roidb

    def rpn_roidb(self):#使用rpn_roidb这种method从imdb中roidb数据
        if int(self._year) == 2007 or self._image_set != 'test':#数据集名称包含2017年且不是test数据集可以用RPN_roidb这种method
            gt_roidb = self.gt_roidb()#使用gt_roidb获得ground truth的roidb,其实就是从XML中解析得到的
            rpn_roidb = self._load_rpn_roidb(gt_roidb)#生成rpn_roidb
            roidb = imdb.merge_roidbs(gt_roidb, rpn_roidb)#将rpn_roidb和gt_roidb组合
        else:
            roidb = self._load_rpn_roidb(None)

        return roidb

    def _load_rpn_roidb(self, gt_roidb):#获得RPN产生的roidb
        filename = self.config['rpn_file']#获取RPN——file文件，它就是RPN网络产生的proposal
        print 'loading {}'.format(filename)
        assert os.path.exists(filename), \
               'rpn data not found at: {}'.format(filename)
        with open(filename, 'rb') as f:
            box_list = cPickle.load(f)#将RPN网络产生proposal的文件加载到box_list文件中
        return self.create_roidb_from_box_list(box_list, gt_roidb) #调用超类imdb的这个函数产生roidb

    def _load_selective_search_roidb(self, gt_roidb):
        filename = os.path.abspath(os.path.join(cfg.DATA_DIR,
                                                'selective_search_data',
                                                self.name + '.mat'))
        assert os.path.exists(filename), \
               'Selective search data not found at: {}'.format(filename)
        raw_data = sio.loadmat(filename)['boxes'].ravel()

        box_list = []
        for i in xrange(raw_data.shape[0]):
            boxes = raw_data[i][:, (1, 0, 3, 2)] - 1
            keep = ds_utils.unique_boxes(boxes)
            boxes = boxes[keep, :]
            keep = ds_utils.filter_small_boxes(boxes, self.config['min_size'])
            boxes = boxes[keep, :]
            box_list.append(boxes)

        return self.create_roidb_from_box_list(box_list, gt_roidb)

    def _load_pascal_annotation(self, index): #解析XML中的信息
        """
        Load image and bounding boxes info from XML file in the PASCAL VOC
        format.
        """
        filename = os.path.join(self._data_path, 'Annotations', index + '.xml')
        tree = ET.parse(filename) #从硬盘导入XML文件，将XML文档解析为ElementTree对象。
        objs = tree.findall('object') #找到所有属于某个tag的element
        if not self.config['use_diff']:
            # Exclude the samples labeled as difficult排除标记为困难的样本
            non_diff_objs = [
                obj for obj in objs if int(obj.find('difficult').text) == 0] #寻找‘difficult’tag中的值为0的obj
            # if len(non_diff_objs) != len(objs):
            #     print 'Removed {} difficult objects'.format(
            #         len(objs) - len(non_diff_objs))
            objs = non_diff_objs
        num_objs = len(objs)

        boxes = np.zeros((num_objs, 4), dtype=np.uint16) #boxes存储坐标，num_objs（样本物体个数） x4
        gt_classes = np.zeros((num_objs), dtype=np.int32) #gt_classes 存储要分的类别，这里的类别数等于num_objs数
        overlaps = np.zeros((num_objs, self.num_classes), dtype=np.float32) #存储重叠率的矩阵，num_obj x num_classes（样本数x类别数）
        # "Seg" area for pascal is just the box area
        seg_areas = np.zeros((num_objs), dtype=np.float32) #h候选框的面积，个数就是Box的个数

        # Load object bounding boxes into a data frame.
    #对每一个objs中的obj进行操作
        for ix, obj in enumerate(objs): #ix是索引
            bbox = obj.find('bndbox')
            # Make pixel indexes 0-based
            x1 = float(bbox.find('xmin').text) - 1 #获取gt的坐标信息
            y1 = float(bbox.find('ymin').text) - 1
            x2 = float(bbox.find('xmax').text) - 1
            y2 = float(bbox.find('ymax').text) - 1
            cls = self._class_to_ind[obj.find('name').text.lower().strip()] #取出当前obj的name，变小写，去除字符串头尾 '/n','/t',' ',然后取出对应字典中的引索值，如aeroplane的cls为1
            boxes[ix, :] = [x1, y1, x2, y2] #将坐标信息存储到boxes列表中，e.g. boxes为([猫1的四个bbox值]，[猫2的四个bbox值]..)
            gt_classes[ix] = cls#e.g. （1,7,10,4..）
            overlaps[ix, cls] = 1.0 #因为这里的Box就是gt，所以重叠率设置为1；这样子其实overlaps就成了一个单位矩阵。e.g. 生成类似与one-hot编码[[0,0,0,0,1,0,0,0,][0,0,0,0,1,0,0,0,]]

            seg_areas[ix] = (x2 - x1 + 1) * (y2 - y1 + 1) #gt的面积

        overlaps = scipy.sparse.csr_matrix(overlaps)#将overlaps稀疏矩阵压缩
    #总结类型：以下key的类型依次为array、array、scipy.sparse.csr.csr_matrix、bool、array
        return {'boxes' : boxes, #返回5个key,boxes存储坐标
                'gt_classes': gt_classes, #存储每个Box对应的类索引
                'gt_overlaps' : overlaps, #共有num_classes（类的个数）行，每一行对应的Box的类索引值为1，其他皆为0，后来被转化为稀疏矩阵；
                'flipped' : False, #表示该图片还未被翻转。
                'seg_areas' : seg_areas}
#roidb的结构是一个包含有5个key的字典

    def _get_comp_id(self):#        self._comp_id = 'comp4'
        comp_id = (self._comp_id + '_' + self._salt if self.config['use_salt']
            else self._comp_id)
        return comp_id

    def _get_voc_results_file_template(self):
        # VOCdevkit/results/VOC2007/Main/_det_test_aeroplane.txt
        filename = self._get_comp_id() + '_det_' + self._image_set + '_{:s}.txt'
        path = os.path.join(
            self._devkit_path,
            'results',
            'VOC' + self._year,
            'Main',
            filename)
        return path

    def _write_voc_results_file(self, all_boxes):
        for cls_ind, cls in enumerate(self.classes):
            if cls == '__background__':
                continue
            print 'Writing {} VOC results file'.format(cls)
            filename = self._get_voc_results_file_template().format(cls)
            with open(filename, 'wt') as f:
                for im_ind, index in enumerate(self.image_index):
                    dets = all_boxes[cls_ind][im_ind]
                    if dets == []:
                        continue
                    # the VOCdevkit expects 1-based indices
                    for k in xrange(dets.shape[0]):
                        f.write('{:s} {:.3f} {:.1f} {:.1f} {:.1f} {:.1f}\n'.
                                format(index, dets[k, -1],
                                       dets[k, 0] + 1, dets[k, 1] + 1,
                                       dets[k, 2] + 1, dets[k, 3] + 1))

    def _do_python_eval(self, output_dir = 'output'):
        annopath = os.path.join(
            self._devkit_path,
            'VOC' + self._year,
            'Annotations',
            '{:s}.xml')
        imagesetfile = os.path.join(
            self._devkit_path,
            'VOC' + self._year,
            'ImageSets',
            'Main',
            self._image_set + '.txt')
        cachedir = os.path.join(self._devkit_path, 'annotations_cache')
        aps = []
        # The PASCAL VOC metric changed in 2010
        use_07_metric = True if int(self._year) < 2010 else False
        print 'VOC07 metric? ' + ('Yes' if use_07_metric else 'No')
        if not os.path.isdir(output_dir):
            os.mkdir(output_dir)
        for i, cls in enumerate(self._classes):
            if cls == '__background__':
                continue
            filename = self._get_voc_results_file_template().format(cls)
            rec, prec, ap = voc_eval(
                filename, annopath, imagesetfile, cls, cachedir, ovthresh=0.5,
                use_07_metric=use_07_metric)
            aps += [ap]
            print('AP for {} = {:.4f}'.format(cls, ap))
            with open(os.path.join(output_dir, cls + '_pr.pkl'), 'w') as f:
                cPickle.dump({'rec': rec, 'prec': prec, 'ap': ap}, f)
        print('Mean AP = {:.4f}'.format(np.mean(aps)))
        print('~~~~~~~~')
        print('Results:')
        for ap in aps:
            print('{:.3f}'.format(ap))
        print('{:.3f}'.format(np.mean(aps)))
        print('~~~~~~~~')
        print('')
        print('--------------------------------------------------------------')
        print('Results computed with the **unofficial** Python eval code.')
        print('Results should be very close to the official MATLAB eval code.')
        print('Recompute with `./tools/reval.py --matlab ...` for your paper.')
        print('-- Thanks, The Management')
        print('--------------------------------------------------------------')

    def _do_matlab_eval(self, output_dir='output'):
        print '-----------------------------------------------------'
        print 'Computing results with the official MATLAB eval code.'
        print '-----------------------------------------------------'
        path = os.path.join(cfg.ROOT_DIR, 'lib', 'datasets',
                            'VOCdevkit-matlab-wrapper')
        cmd = 'cd {} && '.format(path)
        cmd += '{:s} -nodisplay -nodesktop '.format(cfg.MATLAB)
        cmd += '-r "dbstop if error; '
        cmd += 'voc_eval(\'{:s}\',\'{:s}\',\'{:s}\',\'{:s}\'); quit;"' \
               .format(self._devkit_path, self._get_comp_id(),
                       self._image_set, output_dir)
        print('Running:\n{}'.format(cmd))
        status = subprocess.call(cmd, shell=True)

    def evaluate_detections(self, all_boxes, output_dir):
        self._write_voc_results_file(all_boxes)
        self._do_python_eval(output_dir)
        if self.config['matlab_eval']:
            self._do_matlab_eval(output_dir)
        if self.config['cleanup']:
            for cls in self._classes:
                if cls == '__background__':
                    continue
                filename = self._get_voc_results_file_template().format(cls)
                os.remove(filename)

    def competition_mode(self, on):
        if on:
            self.config['use_salt'] = False
            self.config['cleanup'] = False
        else:
            self.config['use_salt'] = True
            self.config['cleanup'] = True

if __name__ == '__main__':
    from datasets.pascal_voc import pascal_voc
    d = pascal_voc('trainval', '2007')
    res = d.roidb
    from IPython import embed; embed()