Late Fusion of Multiple Convolutional Layers for Pedestrian Detection

Abstract : We propose a system design for pedestrian detection by leveraging the power of multiple convolutional layers explicitly. We quantify the effect of different convolutional layers on the detection of pedestrians of varying scales and occlusion level. We show that earlier convolutional layers are better at handling small-scale and partially oc-cluded pedestrians. We take cue from these conclusions and propose a pedestrian detection system design based on Faster-RCNN which leverages multiple convolutional layers by late fusion. In our design, we introduce height-awareness in the loss function to make the network emphasize on pedestrian heights which are misclassified during the training process. The proposed system design achieves a log-average miss-rate of 9.25% on the caltech-reasonable dataset. This is within 1.5% of the current state-of-art approach , while being a more compact system.
Document type :
Conference papers
Complete list of metadatas

https://hal.inria.fr/hal-01926073
Contributor : Ujjwal Ujjwal <>
Submitted on : Sunday, November 18, 2018 - 11:02:26 PM
Last modification on : Tuesday, November 20, 2018 - 1:18:08 AM
Long-term archiving on : Tuesday, February 19, 2019 - 1:04:47 PM

File

Late_Fusion_of_Multiple_Convol...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01926073, version 1

Citation

Ujjwal Ujjwal, Aziz Dziri, Bertrand Leroy, Francois Bremond. Late Fusion of Multiple Convolutional Layers for Pedestrian Detection. 15th IEEE International Conference on Advanced Video and Signal-based Surveillance, Nov 2018, Auckland, New Zealand. ⟨hal-01926073⟩

Share

Metrics

Record views

73

Files downloads

96