Computer vision
Trend 7: Image segmentation, classification, and attribute extraction through AI
Object detection, segmentation, and classification are the building blocks to address complex computer vision challenges. Object detection helps identify an object in the image, forms a rectangular boundary, and creates a bounding box to narrow down the object. Then, image segmentation identifies the object with all curves, lines, and the exact shape. This helps in a more granular and finer identification. This process also helps to establish various insights from images by classifying them, segmenting specific information, and/or extracting any image attributes. Object classification helps classify a particular object into a class or subclass — for example, vehicle classification by type (car, airplane, etc.), and then by brand (Audi, BMW, etc.).
This technique has considerably benefited the health care sector — identifying and narrowing tumor regions and further classifying them as malignant or not.
A large global energy company partnered with Infosys to optimize its cable diagnostics and repair operations to identify faulty cables based on the picture sent from the site, allowing them to take appropriate action. This helped the company to save costs and efforts.
A large global retailer wanted to develop a solution to extract and classify information from digitally scanned product art (SmartArt). The company partnered with Infosys to develop an AI model that can extract information accurately. The information could be further classified as contents, ingredients, instructions, etc. This made the information available on multiple channels for regulatory and compliance purposes.
Computer vision
Trend 8: AI and cloud power video insights
AI's application to videos offers interesting possibilities, such as generating video captions, video highlights, content moderation, brand coverage timings, surveillance, and people/object tracking. For applications like these, cloud computing is necessary for most inference tasks. In fact, object tracking and surveillance are far more powerful in the cloud than on devices, even with new advances in light detection and ranging technology on edge devices such as iPhone. For video processing, the cloud provides scene segmentation and multicamera scene reconstruction; on devices (or the edge — see the next section), only basic segmentation and de-noising capabilities are achieved.
As part of a prestigious global tennis tournament, Infosys extracted various game insights using CV-based algorithms. Event highlights, such as players waving to the crowd, extracting the score from the video feed, recognizing players, and determining the timeframe of a particular advertisement during the live telecast, were created.
Similarly, a large railroad company in the U.S. identified various assets spread across geographies with the help of a streaming video feed from a train-mounted camera.
Other use cases in the CV space include the digitization of KYC form filling, activity and pose recognition in videos, video synthesis, video summarization, and image captioning by leveraging state-of-the-art AI models and techniques such as 3D object detection, generative networks, and singleshot learning.
