Napredna pretraga

Pregled bibliografske jedinice broj: 1011896

Pedestrian Tracking by Probabilistic Data Association and Correspondence Embeddings


Bićanić, Borna; Oršić, Marin; Marković, Ivan; Šegvić, Siniša; Petrović, Ivan
Pedestrian Tracking by Probabilistic Data Association and Correspondence Embeddings // International Conference on Information Fusion (FUSION)
Ottawa, Canada, 2019. str. 1-6 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)


Naslov
Pedestrian Tracking by Probabilistic Data Association and Correspondence Embeddings

Autori
Bićanić, Borna ; Oršić, Marin ; Marković, Ivan ; Šegvić, Siniša ; Petrović, Ivan

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
International Conference on Information Fusion (FUSION) / - , 2019, 1-6

Skup
International Conference on Information Fusion (FUSION)

Mjesto i datum
Ottawa, Canada, 02-05.07.2019.

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
Pedestrian Tracking ; Deep Neural Networks ; JIPDAF ; Multitarget Tracking

Sažetak
This paper studies the interplay between kinematics (position and velocity) and appearance cues for establishing correspondences in multi-target pedestrian tracking. We investigate tracking-by-detection approaches based on a deep learning detector, joint integrated probabilistic data association (JIPDA), and appearance-based tracking of deep correspondence embeddings. We first addressed the fixed-camera setup by fine-tuning a convolutional detector for accurate pedestrian detection and combining it with kinematic-only JIPDA. The resulting submission ranked first on the 3DMOT2015 benchmark. However, in sequences with a moving camera and unknown ego-motion, we achieved the best results by replacing kinematic cues with global nearest neighbor tracking of deep correspondence embeddings. We trained the embeddings by fine-tuning features from the second block of ResNet-18 using angular loss extended by a margin term. We note that integrating deep correspondence embeddings directly in JIPDA did not bring significant improvement. It appears that geometry of deep correspondence embeddings for soft data association needs further investigation in order to obtain the best from both worlds.

Izvorni jezik
Engleski

Znanstvena područja
Elektrotehnika, Računarstvo, Temeljne tehničke znanosti



POVEZANOST RADA


Ustanove
Fakultet elektrotehnike i računarstva, Zagreb