Learning Hybrid Image Templates (HIT) by Information Projection

This paper covers a method of taking images of physical parts which are then preprocessed and compared against CAD generated templates. A pseudo milling operation was performed on discretized points along CAD generated mill paths to create binary image templates. The computer-generated images were then tested against one another as a preliminarily sorting technique. This was done to reduce the number of sorting approaches used, by selecting the most reliable and discerning ones, and discarding the others. To apply the selected sorting methods for comparing CAD generated images and the images of physical parts, a translational and scaling normalization technique was implemented. Rotational variation occurs while scanning physical parts and it was addressed using two different techniques: first by determination of best rotation based on modified-Hausdorff distance (MHD); and second by comparing against all CAD based images for all template rotations. The proposed approach for automated sorting of physical parts was demonstrated by categorizing multiple geometries.

Download Full-text

Eyeglasses detection, location and frame discriminant based on edge information projection

Multimedia Tools and Applications ◽

10.1007/s11042-017-5080-4 ◽

2017 ◽

Vol 77 (12) ◽

pp. 14931-14949 ◽

Cited By ~ 1

Author(s):

Minghua Zhao ◽

Xin Zhang ◽

Zhenghao Shi ◽

Tang Chen ◽

Feifei Zhang

Keyword(s):

Edge Information ◽

Information Projection

Download Full-text

XIMIT – X-Ray Migration Measurement Using Implant Models and Image Templates

Medical Image Computing and Computer-Assisted Intervention – MICCAI 2000 - Lecture Notes in Computer Science ◽

10.1007/978-3-540-40899-4_128 ◽

2000 ◽

pp. 1195-1204

Author(s):

Kathrin Burckhardt ◽

Gábor Székely

Keyword(s):

X Ray ◽

Image Templates

Download Full-text

Neural Network Based Detection of Driver’s Drowsiness

10.36227/techrxiv.14816010 ◽

2021 ◽

Author(s):

Shraddha Bhandarkar ◽

Tanvi Naxane ◽

Sayli Shrungare ◽

Shivani Rajhance

Keyword(s):

Neural Network ◽

Detection System ◽

Traditional Methods ◽

Drowsiness Detection ◽

Proposed Model ◽

Falling Asleep ◽

Image Dataset ◽

Behavioral Aspects ◽

The Face ◽

Image Templates

<div>The primary purpose of this paper was to propose a way to alert sleepy drivers in the act of driving. Most of the traditional methods to detect drowsiness are based on behavioral aspects while some are intrusive and may distract drivers, while some require expensive sensors/hardware. Therefore, in this paper, driver’s drowsiness detection system is developed and implemented to aid drowsy drivers from falling asleep and to prevent accidents. The system takes images from the device as input. Using these image templates, the trained model starts execution and predicts/classifies whether the face of the person in the image is drowsy or alert. The proposed model is able to achieve accuracy of 99.93% using CNN on trained image dataset.</div>

Download Full-text

Multi-Perspective Information Projection: Structuring Hypermedia Information in Engineering Development Projects

Volume 6: 11th Engineering Information Management Conference ◽

10.1115/detc97/eim-3721 ◽

1997 ◽

Author(s):

Srikanth Kannapan ◽

Dean Taylor

Keyword(s):

Ad Hoc ◽

Information Overload ◽

Keyword Search ◽

Work Groups ◽

Specific Work ◽

Engineering Development ◽

Document Models ◽

Shared Information ◽

Information Projection ◽

Distributed Enterprises

Abstract Distributed enterprises are now collaborating closely with their suppliers and partners worldwide while continuously improving their response time to global markets. Such distributed enterprises demand greater levels of support for information sharing and collaborative workspaces among engineering development teams. While the use of Intranets for gathering and sharing information is expanding, the organization of the shared information is usually ad-hoc and is not designed to efficiently serve the diverse information needs of work-groups. Such ad-hoc organizations of information result information overload and hence information users have to resort to (syntactic) keyword search to access information. This paper proposes an Information Projection methodology that integrates the organization of the semantic content of shared information with the organization of syntactic content of documents in a way that is meaningful to specific work-groups so as to reduce information overload. The Information Projection methodology has been applied in three different engineering contexts to varied degrees of sophistication and has shown a significant degree of re-usability, although the information and document models that result are themselves very context specific. The information maps resulting from the context of a MEMS technology development project have also been implemented using basic Web technology (HTML, CGI scripts) and has provided several insights on the implications and limitations of such implementations.

Download Full-text