Repository logo
 

Object Counting with Deep Learning

dc.contributor.advisorStavness, Ian Kent
dc.contributor.committeeMemberMakaroff, Dwight
dc.contributor.committeeMemberRoy, Chanchal
dc.contributor.committeeMemberKo, Seok-Bum
dc.creatorAich, Shubhra
dc.date.accessioned2019-07-04T14:32:02Z
dc.date.available2019-07-04T14:32:02Z
dc.date.created2019-04
dc.date.issued2019-07-04
dc.date.submittedApril 2019
dc.date.updated2019-07-04T14:32:03Z
dc.description.abstractThis thesis explores various empirical aspects of deep learning or convolutional network based models for efficient object counting. First, we train moderately large convolutional networks on comparatively smaller datasets containing few hundred samples from scratch with conventional image processing based data augmentation. Then, we extend this approach for unconstrained, outdoor images using more advanced architectural concepts. Additionally, we propose an efficient, randomized data augmentation strategy based on sub-regional pixel distribution for low-resolution images. Next, the effectiveness of depth-to-space shuffling of feature elements for efficient segmentation is investigated for simpler problems like binary segmentation -- often required in the counting framework. This depth-to-space operation violates the basic assumption of encoder-decoder type of segmentation architectures. Consequently, it helps to train the encoder model as a sparsely connected graph. Nonetheless, we have found comparable accuracy to that of the standard encoder-decoder architectures with our depth-to-space models. After that, the subtleties regarding the lack of localization information in the conventional scalar count loss for one-look models are illustrated. At this point, without using additional annotations, a possible solution is proposed based on the regulation of a network-generated heatmap in the form of a weak, subsidiary loss. The models trained with this auxiliary loss alongside the conventional loss perform much better compared to their baseline counterparts, both qualitatively and quantitatively. Lastly, the intricacies of tiled prediction for high-resolution images are studied in detail, and a simple and effective trick of eliminating the normalization factor in an existing computational block is demonstrated. All of the approaches employed here are thoroughly benchmarked across multiple heterogeneous datasets for object counting against previous, state-of-the-art approaches.
dc.format.mimetypeapplication/pdf
dc.identifier.urihttp://hdl.handle.net/10388/12155
dc.subjectObject counting
dc.subjectDeep learning
dc.subjectOne-look models
dc.subjectHeatmap Regulation
dc.subjectGlobal Sum Pooling
dc.titleObject Counting with Deep Learning
dc.typeThesis
dc.type.materialtext
thesis.degree.departmentComputer Science
thesis.degree.disciplineComputer Science
thesis.degree.grantorUniversity of Saskatchewan
thesis.degree.levelMasters
thesis.degree.nameMaster of Science (M.Sc.)

Files

Original bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
AICH-THESIS-2019.pdf
Size:
28.75 MB
Format:
Adobe Portable Document Format
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
LICENSE.txt
Size:
2.27 KB
Format:
Plain Text
Description: