+ All Categories
Home > Documents > Week 4 : Web-Assisted Object Detection

Week 4 : Web-Assisted Object Detection

Date post: 15-Feb-2016
Category:
Upload: salome
View: 32 times
Download: 0 times
Share this document with a friend
Description:
Week 4 : Web-Assisted Object Detection. Alejandro Torroella & Amir R. zamir. Pre-trained DPM model: Bicycle. Images with bicycles in the frame:. Pre-trained DPM model: Bicycle. Images without bicycles in the frame:. trained DPM model on PASCAL VOC2012 Dataset: Bicycle. - PowerPoint PPT Presentation
Popular Tags:
12
WEEK 4: WEB-ASSISTED OBJECT DETECTION ALEJ AND RO TOR ROELL A & AMIR R. ZAMI R
Transcript
Page 1: Week 4 : Web-Assisted Object Detection

WEEK 4:

WEB-ASSISTED OBJECT

DETECTION

A L E J A N D R O T O R R O E L L A & A MI R R . Z A M

I R

Page 2: Week 4 : Web-Assisted Object Detection

PRE-TRAINED DPM MODEL: BICYCLEImages with bicycles in the frame:

Page 3: Week 4 : Web-Assisted Object Detection

PRE-TRAINED DPM MODEL: BICYCLEImages without bicycles in the frame:

Page 4: Week 4 : Web-Assisted Object Detection

TRAINED DPM MODEL ON PASCAL VOC2012 DATASET: BICYCLE

Images with bicycles in the frame:

Page 5: Week 4 : Web-Assisted Object Detection

TRAINED DPM MODEL ON PASCAL VOC2012 DATASET: BICYCLE

Images without bicycles in the frame:

Page 6: Week 4 : Web-Assisted Object Detection

TRAINED DPM MODEL ON IMAGE-NET DATASET: TRAFFIC LIGHTS

Images with traffic lights in the frame:

Page 7: Week 4 : Web-Assisted Object Detection

TRAINED DPM MODEL ON IMAGE-NET DATASET: TRAFFIC LIGHTS

Images without traffic lights in the frame:

Page 8: Week 4 : Web-Assisted Object Detection

TRAINED DPM MODEL ON STEFFI MORRIS’ DATASET: TRAFFIC LIGHTS

Images with traffic lights in the frame:

Page 9: Week 4 : Web-Assisted Object Detection

TRAINED DPM MODEL ON STEFFI MORRIS’ DATASET: TRAFFIC LIGHTS

Images without traffic lights in the frame:

Page 10: Week 4 : Web-Assisted Object Detection

CONCLUSIONS:• DPM model trained on the Image-Net dataset performed better

than Steffi Morris’ manually annotated dataset.• Likely due to the fact that Steffi’s dataset was much smaller

(~150 vs ~1200)• I believe that both datasets can be better annotated (include

pose) to increase performance.

• DPM model I trained on the VOC2012 dataset performed ever so slightly better than the model pre-trained on the VOC2010 dataset• Makes sense since the VOC2010 dataset is a subset of the

VOC2012 dataset

Page 11: Week 4 : Web-Assisted Object Detection

GIS DATASETS: LOS ANGELES AND D.C.Found GIS data on fire hydrant, street lights, traffic lights and bus stops for the Los Angeles county

Found GIS data for fire hydrants, metro entrances, bus stops, and AM/FM/Cell towers for Washington D.C.

Final choice of dataset to use will depend on DPM results on metro stations, street lights and AM/FM/Cell towers, which I have doubts on how well they can be detected and the quality of the training dataset that can be found on these objects.

Page 12: Week 4 : Web-Assisted Object Detection

THANK YOUFIN.


Recommended