Homepage

This is the homepage of Pascal Mettes.
PhD candidate at the University of Amsterdam.

Latest news

30-03-2017
[Accepted paper] Music-Guided Video Summarization using Quadratic Assignments has been accepted for ICMR 2017.

18-07-2016
[Accepted paper] Spot On: Action Localization from Pointly-Supervised Proposals has been accepted for ECCV 2016.
[Accepted paper] Weakly-supervised recognition, localization and explanation of visual entities has been accepted for the ACM MM 2016 doctorial symposium.
06-05-2016
[Accepted paper] "Featureless: Bypassing Feature Extraction in Action Categorization" is accepted for ICIP 2016.
27-04-2016
[Pre-print available] "Spot On: Action Localization from Pointly-Supervised Proposals" is available online at arXiv.
08-04-2016
[Accepted paper] "Water Detection through Spatio-Temporal Invariant Descriptors" has been accepted for publication by CVIU.
30-03-2016
[Accepted paper] "The ImageNet Shuffle: Reorganized Pre-training for Video Event Detection" is accepted for ICMR 2016.
22-03-2016
Bag-of-Fragments demonstrated at ICT.Open 2016.
23-02-2016
[Pre-print available] "The ImageNet Shuffle: Reorganized Pre-training for Video Event Detection" is available online at arXiv.

03-11-2015
[Pre-print available] "No Spare Parts: Sharing Part Detectors for Image Categorization" is available online at arXiv.
17-09-2015
We finished first in 3 Multimedia Event Detection benchmarks at TRECVID 2015.
14-09-2015
Bag-of-Fragments demonstrated at NCCV 2016.
24-03-2015
No Spare Parts demonstrated at ICT.Open 2016.
16-03-2015
[Accepted paper] "Bag-of-Fragments: Selecting and Encoding Video Fragments for Event Detection and Recounting" has been accepted for ICMR.

21-07-2014
[Accepted paper] "Nature Conservation Drones for Automatic Localization and Counting of Animals" has been accepted for a workshop at ECCV.

Data

Spot On: Action Localization from Pointly-Supervised Proposals

Hollywood2Tubes video annotations (407 kb).

Please cite our work when using these annotations in your work.

The ImageNet Shuffle: Reorganized Pre-training for Video Event Detection

Caffe model	Prototext	ImageNet IDs	Number of classes
Bottom-up [4k] model (92 MB)	Bottom-up [4k] deploy	Bottom-up [4k] ids	4,437
Bottom-up [8k] model (136 MB)	Bottom-up [8k] deploy	Bottom-up [8k] ids	8,201
Bottom-up [13k] model (192 MB)	Bottom-up [13k] deploy	Bottom-up [13k] ids	12,988
Top-down [4k] model (87 MB)	Top-down [4k] deploy	Top-down [4k] ids	4,000