I am using the Coco eval script for my project. I have modified the demo provided here. The commands I use are basically those: cocoGt = COCO(gt_json) cocoDt = cocoGt.loadRes(results_json) imgIds = sorted(cocoGt.getImgIds()) # running evaluation cocoEval = COCOeval(cocoGt, cocoDt, annType) # cocoEval.params.catIds = [1] # person id : 1 cocoEval.params.imgIds = imgIds cocoEval.evaluate() cocoEval.accumulate() # in this step the results are printed cocoEval.summarize() everything works fine except for a display of the results in the console. The display opccurs when cocoEval.accumulate() is executed. The display is something similar to this one: DONE (t=0.52s). Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.617 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.945 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.688 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.533 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.626 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.869 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.111 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.539 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.717 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.658 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.722 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.925 and it is displayed each time the prompt is used (if on debug or interactive mode that means in EVERY command!). When launched in a script the display is only occurring once but still it shouldn't happen. After investigating the code I found that the printing occurs in line 411 of cocoeval.py. The inexplicable issue is that the next line which cause this console printing is just a dictionary definition: self.eval = { 'params': p, 'counts': [T, R, K, A, M], 'date': datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S'), 'precision': precision, 'recall': recall, 'scores': scores, } The even funnier thing is that if the precision field is empty the printing stops from happening! For example the following dict does not produce anything in the console (of course it crushes after a while and the script does not work correctly). self.eval = { 'params': p, 'counts': [T, R, K, A, M], 'date': datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S'), 'recall': recall, 'scores': scores, } Similarly, if I include a line like cocoEval.eval = {} # this stop the result printing for some reason after cocoEval.summarize() the printing also stops (it has occurred once of course already). precision is an ordinary numpy array also. The questions I have are: a) Why the inclusion of this specific field precision in the cocoEval.eval dict produce this printing change? b) How can it be stopped from occurring? Continue reading...
