The Opportunities and Challenges of Putting the Latest Computer Vision and Deep Learning Algorithms to Work
5 likes1,091 views
The document outlines Albert Y. C. Chen's professional journey and experiences in computer vision and deep learning, highlighting his contributions to the field through numerous publications and innovations. It discusses the evolution of computer vision technologies, the importance of adapting business models in rapidly changing markets, and the rise of startups in driving innovation. Additionally, it addresses the challenges faced in technological advancements and the need for collaboration between academia and industry to foster effective innovation.
![• Final inspection cells
• Robot guidance and
checking orientation of
components
• Packaging Inspection
• Medical vial inspection
• Food pack checks
• Verifying engineered
components[5]
• Wafer Dicing
• Reading of Serial
Numbers
• Inspection of Saw
Blades
• Inspection of Ball Grid
Arrays (BGAs)
• Surface Inspection
• Measuring of Spark
Plugs
• Molding Flash Detection
• Inspection of Punched
Sheets
• 3D Plane
Reconstruction with
Stereo
• Pose Verification of
Resistors
• Classification of Non-
Woven Fabrics
1970s-now: Machine Vision
for Industrial Inspection
• Automated Train
Examiner (ATEx)
Systems
• Automatic PCB
inspection
• Wood quality
inspection
• Final inspection of
sub-assemblies
• Engine part inspection
• Label inspection on
products
• Checking medical
devices for defects](https://image.slidesharecdn.com/20161027sinicaopportunitiesandchallengesforapplyingthelatestcomputervisionanddeeplearningresearchtow-170518085819/85/The-Opportunities-and-Challenges-of-Putting-the-Latest-Computer-Vision-and-Deep-Learning-Algorithms-to-Work-59-320.jpg)
The Opportunities and Challenges of Putting the Latest Computer Vision and Deep Learning Algorithms to Work
- 1. Putting the latest Computer Vision and Deep Learning algorithms to work The Opportunities and Challenges Albert Y. C. Chen, Ph.D. Vice President, R&D Viscovery
- 2. Albert Y. C. Chen, Ph.D. • Experience 2017-present: Vice President of R&D at Viscovery 2016-2017: Chief Scientist at Viscovery 2015: Principal Scientist @ Nervve Technologies 2013-2014 Computer Vision Scientist @ Tandent Vision 2011-2012 @ GE Global Research • Education Ph.D. in Computer Science, SUNY-Buffalo M.S. in Computer Science, NTNU B.S. in Computer Science, NTHU • Some random things about me… SUNY Excellence in Teaching Award, 2010. Some rapid promotions, some failed startups, some patents, some papers…
- 3. 1. W.Wu,A.Y. C. Chen, L. Zhao, and J. J. Corso. Brain tumor detection and segmentation in a CRF framework with pixel-wise affinity and superpixel-level features. International Journal of Computer Assisted Radiology and Surgery, 2015. 2. S. N. Lim,A.Y. C. Chen and X.Yang. Parameter Inference Engine (PIE) on the Pareto Front. In Proceedings of International Conference of Machine Learning,Auto ML Workshop, 2014. 3. A.Y. C. Chen, S.Whitt, C. Xu, and J. J. Corso. Hierarchical supervoxel fusion for robust pixel label propagation in videos. In Submission to ACM Multimedia, 2013. 4. A.Y.C. Chen and J.J. Corso.Temporally consistent multi-class video-object segmentation with the video graph-shifts algorithm. In Proceedings of IEEE Workshop on Applications of ComputerVision, 2011. 5. D.R. Schlegel,A.Y.C. Chen, C. Xiong, J.A. Delmerico, and J.J. Corso. Airtouch: Interacting with computer systems at a distance. In Proceedings of IEEE Workshop on Applications of ComputerVision, 2011. 6. A.Y.C. Chen and J.J. Corso. On the effects of normalization in adaptive MRF Hierarchies. In Proceedings of International Symposium CompIMAGE, 2010. 7. A.Y.C. Chen and J.J. Corso. Propagating multi-class pixel labels throughout video frames. In Proceedings of IEEE Western NewYork Image Processing Workshop, 2010. 8. A.Y. C. Chen and J. J. Corso. On the effects of normalization in adaptive MRF Hierarchies. Computational Modeling of Objects Represented in Images, pages 275–286, 2010. 9. Y.Tao, L. Lu, M. Dewan,A.Y. C. Chen, J. J. Corso, J. Xuan, M. Salganicoff, and A. Krishnan. Multi-level ground glass nodule detection and segmentation in ct lung images. Medical Image Computing and Computer-Assisted Intervention, 2009. 10. A.Y.C. Chen, J.J. Corso, and L.Wang. Hops: Efficient region labeling using higher order proxy neighborhoods. In Proceedings of IEEE International Conference on Pattern Recognition, 2008.
- 4. Some work done before I caught the startup fever Freestyle Sketching Stage AirTouch waits in background for the initialization signal Initialize Terminate Output image database Start: Results CBIR query Airtouch HCI interface for Content-based Image Retrieval
- 5. Interactive Segmentation & Classification • Segmentation then classification: • computationally more efficient, • results in much higher classification accuracy. • Pioneered the “pixel label propagation” field. • First to utilize superpixels and supervoxels for the task. FG Traditional Spatial Propagation Pixel label map Label a subset of pixels BG Spatio-temporal Propagation time
- 6. Image/Video Object Recognition and Content Understanding approaches person carries gives recieves Ontology object Person 1 Person 1Person 2 High-Level Mid-Level approach activity receives gives carries activity activity activity Time Reasoning x x x Low-Level x x x x
- 7. Learning and Adapting Optimal Classifier Parameters subspace B subspace A subspace C Image-level feature space priors Patch-level feature space posterior probability suggest optimal parameter configuration
- 8. Graphical Models and Stochastic Optimization A (a) The space-time volume of a video showing the objects (A--F) and their appearing time-span. space time A B C D E F B E F C D (b) The temporal relationship graph. An edge between two vertices mean that the two objects overlap in time. (c) The goal is: cover all objects with the smallest number of "ground truth key frames". space time A B C D E F key 1 key 2 A B E F C D (d) This translates to: iteratively solving the max clique problem until all vertices belong to a clique. A B E F C D key 2 key 1 frame t-1 frame t layer n layer n layer n+1 layer n+1 Temporal Shift Shift µ
- 9. Medical Imaging and Geospatial Imaging GNN detection and segmentation in Lung CT geospatial imaging: building detection Brain tumor detection and segmentation in MR images.
- 10. Why Risk to Innovate? • Good business model NEVER last forever. • Average “shelf life” on S&P 500: 20 years. • 100-year old companies constantly reinvent themselves every 10-20 years • Startups contribute to 20% of USA’s GDP.
- 11. The Death of a Good Business Model • Foxconn 20 year revenue v.s. net profit (now at 5%)
- 12. What do 100 year old corporations do? GE Schenectady, 1896
- 13. History of change at GE • 1886: one of the 12 original companies on the Dow Jone Industrial Average (also the only one remaining). • 1889: lightbulbs • 1919: radios • 1927: TV • 1941: jet engine • 1960: nuclear power • 1971: room AC units • 1995: MRI
- 14. History of change at IBM • 1960s: mainframe computer • 1980s: personal computer • 2000s: integrated solutions • 2020s: AI, Watson
- 15. How about the leading Semiconductor companies?
- 16. NVidia reinventing itself —2 times in 20 years
- 17. “Bad money drives out good” in the desktop GPU market
- 18. The rise of mobile computing, and how NVidia missed the boat!
- 19. NVidia’s Tegra mobile processors never took off then, the market saturated…
- 20. NVidia not just survived. NVidia is thriving!
- 21. Meet the new NVidia: Deep Learning, Deep Learning, and still, Deep Learning
- 23. The king is dead, long live the king!
- 24. Now, again, do we want to do OEM/ODM forever? Optimizing an old business model is just delaying its eventual death.
- 25. Startups • A company, partnership, or temporary organization designed to search for a new, repeatable and scalable business model.
- 26. Your Idea • Are you passionate about it? • Is it disruptive enough? • What is your business plan? • What is it? • Can it make money? • What is the future of the idea? • What is your competitive advantage? • How do you build up your entry barrier?
- 27. A minimal startup team • A hacker • A hustler • A hipster
- 29. Startup Timeline
- 30. Prototype • Hack out a prototype • Spend 2-10 weeks max. • Investors are much more likely to fund you if you have a minimal initial version of your idea. • Hackathons are a good place to start. • Iteratively improve the prototype
- 31. Money!
- 33. Buildup your entry barrier! • Market (users) • Speed • Team • Technology
- 34. Building entry barrier with Technology!!
- 35. Computer Vision, it can’t be that hard, right?
- 36. Brief History Marvin Minsky “In 1966, Minsky hired a first-year undergraduate student and assigned him a problem to solve over the summer: connect a television camera to a computer and get the machine to describe what it sees.” Gerald Sussman The student never worked on Computer Vision problems again.
- 37. Brief History • 1960’s: interpretation of synthetic worlds • 1970’s: some progress on interpreting selected images • 1980’s: ANNs come and go; shift toward geometry and increased mathematical rigor • 1990’s: face recognition; statistical analysis in vogue • 2000’s: broader recognition; large annotated datasets available; video processing starts Guzman ‘68 Ohta Kanade ‘78 Turk and Pentland ‘91
- 38. What’s in our arsenal? • Image filters • Feature descriptors • Classifiers
- 40. Filters: sharpen
- 41. Filters: edge
- 43. Features:
- 45. Features: Laplacian of Gaussian (LoG; scale detection)
- 46. Features: Orientation How to compute the rotation? Create edge orientation histogram and find peak.
- 47. Features: SIFT
- 48. Features: SIFT
- 49. Features: Gabor
- 50. Classifiers: SVM
- 53. Classifiers: Deformable Parts Model (DPM)
- 54. Classifiers: Deep Neural Network
- 55. What alg. should I use then? • How much data do we have? • What objects are we trying to detect? • For example, Google’s DNN trained with 11k images over 20 classes in 2013 doesn’t always beat DPM. 0 0.15 0.3 0.45 0.6 aero bike bird boat bottle bus car cat chair cow 0 0.15 0.3 0.45 0.6 dog horse m-bike person plant sheep sofa table train TV D N N D P M
- 56. ML alg. and their Applications • Deep Learning • Markovian/ Bayesian • Feature Matching • Other ML methods
- 57. Meta-Learning • Different use cases calls for different ML algorithms. • Meta-Learning: learning how to learn. • Requires plenty of domain-specific know-how.
- 59. • Final inspection cells • Robot guidance and checking orientation of components • Packaging Inspection • Medical vial inspection • Food pack checks • Verifying engineered components[5] • Wafer Dicing • Reading of Serial Numbers • Inspection of Saw Blades • Inspection of Ball Grid Arrays (BGAs) • Surface Inspection • Measuring of Spark Plugs • Molding Flash Detection • Inspection of Punched Sheets • 3D Plane Reconstruction with Stereo • Pose Verification of Resistors • Classification of Non- Woven Fabrics 1970s-now: Machine Vision for Industrial Inspection • Automated Train Examiner (ATEx) Systems • Automatic PCB inspection • Wood quality inspection • Final inspection of sub-assemblies • Engine part inspection • Label inspection on products • Checking medical devices for defects
- 60. Industrial Inspection: turbofan jet engine blade maintenance • Some seemingly daunting machine vision tasks actually works with relatively simple image processing algorithms.
- 61. Industrial Inspection: Cognex Omniview
- 62. Industrial Inspection: Cognex Omniview
- 64. License Plate Readers with Text Detection and Neural Networks
- 65. Biometrics
- 67. Face Recognition (1990s-now) • Face Detection (Viola and Jones, 2001) • Face Verification (1:1) v.s. Identification (1:N)
- 68. Face Verification and Identification, Labeled Faces in the Wild (LFW) Recognition Accuracy: • 1 to 1: 99%+ • 1 to 100: 90% • 1 to 10,000: 50%-70%. • 1 to 1M: 30%. LFW dataset, common FN↑, FP↓
- 70. Sports—NFL first down line (1995-now)
- 71. Sports—NFL first down line minus equals
- 72. 3D Reconstruction (As old as CV; became practical since SIFT)
- 73. 3D Reconstruction with Feature Matching, Structure from Motion
- 74. 3D Reconstruction with Feature Matching, Structure from Motion
- 75. Image Panoramas (1980s - now)
- 76. Solving Panorama Problem with Markov Random Fields Input:
- 77. Solving Panorama Problem with Markov Random Fields Input:
- 78. Solving Panorama Problem with Markov Random Fields Input:
- 79. Solving Panorama Problem with Markov Random Fields Input:
- 80. Solving Panorama Problem with Markov Random Fields Input:
- 81. Solving Panorama Problem with Markov Random Fields Input:
- 82. Solving Panorama Problem with Markov Random Fields Input:
- 83. Solving Panorama Problem with Markov Random Fields
- 84. Solving Panorama Problem with Markov Random Fields ICM (Iterated Conditional Modes), 1986
- 85. Solving Panorama Problem with Markov Random Fields Belief Propagation (1980-2000)
- 86. Solving Panorama Problem with Markov Random Fields Graph-Cuts (alpha expansion), 2001
- 87. Photosynthesis
- 88. Solving Photosynthesis Problems with Alpha-matting (2000s-now)
- 89. Object Detection & Classification state-of-the-art • ImageNet Large Scale Visual Recognition Challenge (ILSVRC) • 1000+ classes, 1.2M images. 0 0.125 0.25 0.375 0.5 11 12 13 14 11 12 13 14 classification error classification +localization error
- 90. Image Scene Classification • MIT Places 401 dataset. • top-5 accuracy rates >80%.
- 93. 2005 winner, Stanley (Stanford), 3mph through desert
- 96. 2007 winner, Boss (CMU), 13mpg through the city
- 97. Self Driving Cadillac, US congressman to airport, 2013
- 98. Google Self Driving Car, 2015
- 100. Google Self-Driving Car, 2016
- 101. NVidia Self Driving Car, 2016
- 102. How did we come this far? Race car drivers know the trick
- 103. Focus on Free Space / Drivable Area, not Obstacles!
- 105. Structure from X, Floored
- 106. Structure from X, PIX4D
- 107. Object Recognition Blue River Technology
- 109. IMRSV
- 111. Other Applications in Business Intelligence • Measure brand exposure. • Measure sponsorship effectiveness. • Loss prevention and retail layout optimization.
- 113. Angel.co
- 114. My humble attempts at putting the latest Computer Vision algorithms to work
- 115. Intrinsic Imaging at Tandent Vision Science Computer Vision would be half-solved without shadows! LightOriginal Image Surface
- 116. Tandent Lightbrush Video Tutorial for Tandent Lightbrush: https://vimeo.com/47009123
- 118. Issues • Highly anticipated, highly acclaimed, but small crowd at $500 a license. • Adobe Photoshop monopoly and the “not invented here” syndrome. • Adobe’s arch-rival, Corel (Corel Draw, Paint Shop Pro, Ulead PhotoImpact) was DYING and asked too much from the botched deal.
- 119. Have fun scribbling out your shadows in photoshop! Poor Bob from Adobe wasted 9 minutes removing just 1 shadow
- 120. Intrinsic Imaging for improving the RGB signal in autonomous driving
- 122. Retrospect • 20 researchers burned 25 million in 8 years; investors got 50 patents in return, period. • Overestimated the total addressable market size, in a market with existing monopoly. • Many missed opportunities. Counterexample of the lean startup model.
- 123. Some SfM, SLAM startups
- 124. Satellite/Aerial Imagery Analysis • 40cm resolution at 30fps for 90 sec for any location on earth. • One LEO satellite revisits any place on Earth every 3 days. • Need 24 satellites to revisit any place on Earth every 3 hours.
- 125. Challenges for Single satellite depth estimation and 3D reconstruction • At 30fps, a LEO satellite travels 250m between two consecutive frames —> theoretically sufficient for cm-level depth estimation. • Sources of Noise: • Camera distortions • Atmospheric Disturbance • Ground vegetation • Sub-pixel sampling noise 1 2
- 126. What happened? • B2B customers takes too long to strike deals. • Google ate us alive in just 3 months, while we were still pitching for VC-funding with our prototype.
- 127. Visual Search at Nervve
- 128. Retrospect • Growth pains expanding from intelligence community clients to advertisement clients. • Forming the right team of engineers and researchers and moving at the right pace. • For any Computer Vision/Machine Learning company: • Researchers that cannot program—> OUT • Engineers that don’t know math —> OUT
- 129. Visual Search, Simply Smarter
- 130. Once in a lifetime opportunity in China’s video streaming market
- 131. What do we need? Face Motion Image scene Text Audio Object Semantics
- 132. Viscovery VDS (Video Discovery Service)
- 133. Viscovery VDS (Video Discovery Service)
- 134. Viscovery VDS (Video Discovery Service)
- 135. Challenges Encountered Along the Way • From Product Recognition in Images, to Face, Logo, Object, Scene recognition in Videos. • Number of Categories • Recognition Accuracy • Recognition Speed • System Architecture • Business Model
- 136. Viscovery’s Edge • Market: first mover’s advantage in China’s video streaming market. • Speed: we built the whole VDS thing in a few months! • Team: You! Seriously! • Technology: • Depth • Breadth • Cloud • Customizability • Self-Learning
- 137. Life is not all rosy at startups • High Risk, High Pressure, High Uncertainty! • Resources are scarce, but you MUST DELIVER! • Forming your all-star team is not that easy… • Focus, and persistence.
- 138. What can Taiwan’s academia do to help bridge the gap? HMM….
- 139. Academia IndustryGeneral Public reputation and policy support improved living standards students opportunity well-trained graduates grants and collaborations A healthy cycle
- 140. Academia IndustryGeneral Public unsupportive policies stagnant wages useless education unemployable graduates A vicious cycle no grants no students
- 141. Where should we start? Maybe with a few more stories.
- 142. Where should we start? Maybe with a few more stories.
- 143. Where should we start? Maybe with a few more stories.
- 144. The Goldilocks zone of innovation
- 145. The Goldilocks zone of innovation Business Relevance Academic Relevance plentiful resources; hierarchical organization lack of resources; responsive organization traditional corporations talking “innovation” corporate research startups struggling to survive academic spinoffs MSR
Editor's Notes
- #27: 翟本橋：never worked a single day in my life example: Tivo disrupts TV market / creates DVR market example: Facebook, Twitter disrupt online social networking example: FourSquare creates location-based "check in" ad market
- #125: Agriculture Health Monitoring, Humanitarian Aid, Insurance Modeling, Oil Storage Monitoring, Natural Disaster Response, Oil&Gas Infrastructure Monitoring Financial Trading Intelligence, Mining and Logging Monitoring, Maritime Monitoring
- #126: At 681km altitude, orbit speed 27065 km/h = 451 km/ minute = 7.52 km / sec = 250m / frame @ 30fps.