kamkyu94/fast_online_mtmct

Fast online multi-target multi-camera tracking for vehicles

fast_online_mtmct's Issues

The IDF1 value is very low

Hello authors, I have tested using your code on the S02 dataset of AIC22, the detector used is yolov7-e6, I successfully got the mtmc_resnet50_ibn_a_gap.txt file, but when I evaluated the trace results using the eval.py file in the AIC22 eval folder, I got Very low IDF1 values where the IDP values are reasonable but the IDR values are very unreasonable, I am confused about this, I would like to ask you how you evaluated the final trace resolution, looking forward to your reply!

How should I obtain the AIC19 dataset？

How to do Temporal alignment？

` temp_align = {}
for cam in cams:
temp_align[cam] = {}
for i in range(0, np.max(f_nums) + 1):
# Default
temp_align[cam][i] = 0

        # Set for each camera
        if cam == 'c006':
            temp_align[cam][i] = i

        elif cam == 'c007':
            if i <= 1037:
                temp_align[cam][i] = i + 1
            elif 1040 <= i <= 1309:
                temp_align[cam][i] = i - 1
            elif 1320 <= i <= 1339:
                temp_align[cam][i] = i - 11
            elif 1350 <= i <= 1379:
                temp_align[cam][i] = i - 21
            elif 1400 <= i <= 1449:
                temp_align[cam][i] = i - 41
            elif 1460 <= i <= 1499:
                temp_align[cam][i] = i - 51
            elif 1510 <= i <= 1537:
                temp_align[cam][i] = i - 61
            elif 1540 <= i <= 1542:
                temp_align[cam][i] = i - 63
            elif 1560 <= i <= 1609:
                temp_align[cam][i] = i - 80
            elif 1620 <= i <= 1639:
                temp_align[cam][i] = i - 90
            elif 1650 <= i <= 1864:
                temp_align[cam][i] = i - 100
            elif 1870 <= i <= 1893:
                temp_align[cam][i] = i - 105
            elif 1901 <= i <= 1920:
                temp_align[cam][i] = i - 112
            elif 1927 <= i <= 1933:
                temp_align[cam][i] = i - 118
            elif 1940 <= i <= 1989:
                temp_align[cam][i] = i - 124
            elif 2000 <= i <= 2049:
                temp_align[cam][i] = i - 134
            elif 2060 <= i:
                temp_align[cam][i] = i - 144

        elif cam == 'c008':
            if 7 <= i <= 421:
                temp_align[cam][i] = i - 6
            elif 439 <= i <= 472:
                temp_align[cam][i] = i - 23
            elif 479 <= i <= 548:
                temp_align[cam][i] = i - 29
            elif 603 <= i <= 685:
                temp_align[cam][i] = i - 83
            elif 728 <= i <= 925:
                temp_align[cam][i] = i - 125
            elif 934 <= i <= 1397:
                temp_align[cam][i] = i - 133
            elif 1401 <= i <= 1612:
                temp_align[cam][i] = i - 136
            elif 1621 <= i <= 1752:
                temp_align[cam][i] = i - 144
            elif 1763 <= i <= 1920:
                temp_align[cam][i] = i - 154
            elif 1958 <= i:
                temp_align[cam][i] = i - 191

        elif cam == 'c009':
            temp_align[cam][i] = i - 9`

Hello Dear Author, I would like to ask you some questions. I am trying to use your code on the S06 dataset of AIC22_Track1_MTMC_Tracking, but I am experiencing problems with temporal alignment. I would like to know how you implemented the time alignment, especially in terms of determining the time alignment of the image frames. I'm not very clear on the methodology and implementation of this, so I was hoping you could provide some guidance or explanation. Thank you very much for your help!