6 Optical Flow Jobs

We are looking for a Computer Vision Engineer to contribute to the development of AI-driven video analytics solutions tailored for extracting detailed cricket data from various sources such as match footage, live broadcasts, and player tracking systems. In this role, you will be utilizing deep learning techniques, pose estimation algorithms, and image processing methodologies to scrutinize player movements, shot selection, bowling mechanics, and overall match dynamics. Your responsibilities will include: - Developing computer vision models geared towards player tracking, ball trajectory prediction, and shot classification within the realm of cricket analytics. - Extracting real-time insights from video streams, encompassing player positions, footwork analysis, and batting/bowling techniques. - Implementing pose estimation techniques (such as OpenPose, MediaPipe, DeepLabCut) to dissect player body movements. - Employing action recognition models to categorize batting strokes, bowling variations, and fielding plays. You will also be involved in: - Applying various deep learning tools like CNNs, Transformers, YOLO, Mask R-CNN, and Optical Flow for cricket action recognition. - Developing segmentation models to differentiate between players, umpires, and key match elements like the ball, stumps, and pitch zones. - Utilizing GANs and super-resolution techniques to enhance the quality of low-resolution video footage. - Constructing OCR systems for Optical Character Recognition to extract scoreboards and match statistics from live broadcasts. Additional responsibilities include: - Implementing real-time video processing pipelines using OpenCV, FFmpeg, and GStreamer. - Working with multi-angle TV footage and drone feeds to enrich match analysis. - Developing algorithms to ensure low-latency real-time detection of ball speed, swing, and spin. You will also be required to: - Create automated annotation tools for labeling cricket match videos. - Train supervised and unsupervised learning models using cricket datasets. - Collaborate with sports analysts to validate and refine model predictions. Furthermore, your duties will involve: - Deploying computer vision models on edge devices, cloud platforms, and real-time streaming systems. - Optimizing models for high accuracy and low latency in live cricket broadcasts. - Conducting A/B testing and continuous monitoring to enhance model performance. You will also be expected to: - Perform camera calibration for distortion correction and real-world coordinate mapping. - Implement homography transformation and perspective correction for multi-angle alignment. - Apply video trigonometry techniques to calculate player distances, ball speed, angles, and movement trajectories. - Work with multi-camera synchronization and stereo vision for depth estimation in cricket analytics. About the Company: Lifease Solutions LLP believes in the power of design and technology to address challenges and transform ideas into reality. As a prominent provider of software solutions and services, Lifease Solutions is dedicated to delivering innovative and high-quality solutions that drive growth and value for businesses. Headquartered in Noida, India, we specialize in the finance, sports, and capital market sectors, earning the trust of companies worldwide. Our track record showcases our capability to turn minor projects into major successes, and we continuously seek opportunities to help our clients optimize their IT investments.,

As a member of ClearTrail, you will be part of a team where work transcends beyond being just a job. Our mission is to create solutions that empower those committed to ensuring the safety of individuals, locations, and communities. With over 23 years of experience, law enforcement and federal agencies worldwide rely on ClearTrail as their dedicated partner in protecting nations and enhancing lives. We are at the forefront of shaping the future of intelligence gathering through the development of artificial intelligence and machine learning-based solutions for lawful interception and communication analytics. Our focus is on addressing the world's most complex challenges by leveraging cutting-edge technologies. Your primary responsibility will involve developing and training machine learning models that integrate drone, satellite, and real-time video data to achieve advanced visual comprehension. Key tasks include building ML/DL models for image and video classification, segmentation, and spatiotemporal analysis, as well as creating fusion methods to align and merge satellite, drone, and ground footage. You will be expected to apply various techniques such as multi-scale learning, temporal modeling (e.g., 3D CNNs, transformers), and sensor fusion, along with conducting model evaluation and optimization across diverse data sources. To excel in this role, you must possess strong Python programming skills and proficiency in data manipulation using libraries like NumPy, Pandas, and SciPy. Additionally, you should be well-versed in vision/model libraries such as MMDetection, Detectron2, and OpenCV, and have experience in deep learning frameworks like PyTorch and TensorFlow, particularly in video models like TimeSformer, SlowFast, I3D, and Yolo. Familiarity with satellite/drone geospatial fusion, GPS/IMU metadata, scene alignment, time series analysis, motion tracking, and optical flow techniques will be advantageous. Knowledge of remote sensing libraries like Rasterio, GDAL, GeoPandas, and xarray will also be beneficial for this role. Join us at ClearTrail and be part of a dynamic team dedicated to creating innovative solutions that make a real difference in the world of intelligence and security.,

The main objectives of a Drone Systems Engineer include multirotor design and development using design software, UAV assembly, and flight testing of multirotor UAVs. You should have in-depth knowledge of ground control software, expertise in PID tuning, and advanced parameter configuration. Additionally, you will be responsible for subsystem integration such as LiDAR, optical flow, RTK/PPK, and dual GPS. Building strong customer relationships and rapport to effectively fulfill customer needs is also a key aspect of this role. This is a full-time position with benefits such as paid sick time and provident fund. The work schedule is during the day shift, and the work location is in person.,

Vimaan is looking to onboard multiple Machine Learning Engineers in Bengaluru, India to drive the development of computer vision and machine learning algorithms to power our cutting-edge wall-to-wall warehouse inventory tracking and verification platform. This is a unique opportunity to exploit a treasure cove of unseen real-world data coming from a multi-camera perception system and develop large-scale computer vision and deep learning models to build a product that creates a disproportionate value for the warehouse industry. The role involves hands-on CV/ML software development and deployment from understanding the product requirements, defining Computer Vision functional specs to designing, developing, and deploying CV/ML models in production at scale. The ideal candidate for the Machine Learning Engineer position should have an MS in computer vision, machine learning, AI, applied mathematics, data science, or related technical fields, or a BS with 3+ years of experience in Computer Vision/Machine Learning. They should possess hands-on experience in developing new learning algorithms for computer vision tasks such as object detection, object tracking, instance segmentation, activity detection, depth estimation, optical flow, multi-view geometry, domain adaptation, adversarial and generative models, as well as representational learning with varying amounts of data. Knowledge of current deep learning literature and the mathematical foundations of machine learning is essential. Experience with popular object detection frameworks such as YOLO, SSD, or Faster R-CNN is considered a plus. The candidate should have the ability to train and debug deep learning systems, gain deep insights into data characteristics, and map those to appropriate model architectures. Experience working with inputs from multiple cameras and input modes is advantageous, as is experience in AI Infrastructure, Machine Learning Accelerators, On-Device Optimization, and training and deploying deep learning models on GPU-accelerated platforms. Strong programming skills with Python and experience with ML/DL frameworks like Tensorflow, Pytorch, etc., are required. Prior experience in deploying machine learning models in production environments, working with cloud platforms (e.g., AWS, Azure, Google Cloud), and familiarity with data pre-processing, augmentation, and visualization tools and libraries are necessary. The ideal candidate should be highly motivated, passionate, possess a strong work ethic, and be able to work effectively in a team or independently under supervision in a matrix management environment. Effective communication skills, problem-solving abilities, attention to detail, and a passion for staying at the forefront of technology advancements in machine learning and computer vision are key attributes. They should be able to work in a fast-paced, high-pressure startup environment, adapt to rapidly changing requirements, and convey complex technical concepts to non-technical stakeholders. The Machine Learning Engineer will be responsible for researching, designing, and developing machine learning algorithms and models for various tasks of detection, recognition, and classification for warehouse inventory management. They will implement and optimize deep learning architectures, explore techniques like transfer learning and data augmentation, guide annotation teams, curate and pre-process annotated datasets, collaborate with MLOps for integrating machine learning models into production systems, and conduct thorough performance analysis and evaluation of models using appropriate metrics and tools. It is essential for the Machine Learning Engineer to stay up-to-date with the latest advancements in machine learning and computer vision research and integrate relevant findings into the solutions developed at Vimaan. ABOUT VIMAAN Headquartered in Silicon Valley, with team members around the world, Vimaan is comprised of computer vision and hardware technologists, as well as warehousing domain experts with a successful history in technology startups. Vimaan's primary mission is to deliver computer vision and machine learning solutions to solve long-standing inventory visibility, accuracy, and quality challenges in the supply chain.,

You will be working at ClearTrail, where the focus is on developing solutions to empower organizations dedicated to ensuring the safety of their people, places, and communities. With over 23 years of experience, ClearTrail has been a trusted partner for law enforcement and federal agencies worldwide in safeguarding nations and enriching lives. As part of the team, you will contribute to the future of intelligence gathering by creating artificial intelligence and machine learning-based lawful interception and communication analytics solutions to address some of the world's most challenging problems. Your role will involve developing and training machine learning models that leverage drone, satellite, and real-time video data for complex visual understanding tasks. You should have a minimum of 10-12 years of relevant experience for this position. Key responsibilities of the role include: - Building machine learning and deep learning models for image and video classification, segmentation, and spatiotemporal analysis. - Developing fusion methods to align and combine data from satellite, drone, and ground footage. - Applying advanced techniques such as multi-scale learning, temporal modeling (e.g., 3D CNNs, transformers), and sensor fusion. - Conducting model evaluation and optimization across various data sources. Essential skills for this role include: - Strong proficiency in Python coding and data manipulation using libraries such as NumPy, Pandas, and SciPy. - Expertise in vision/model libraries like MMDetection, Detectron2, and OpenCV. - Deep learning experience with PyTorch, TensorFlow, and video models such as TimeSformer, SlowFast, I3D, and Yolo. - Knowledge of satellite/drone geospatial fusion, GPS/IMU metadata, and scene alignment. - Familiarity with time series analysis, motion tracking, and optical flow using tools like RAFT and FlowNet. - Understanding of remote sensing libraries such as Rasterio, GDAL, GeoPandas, and xarray. If you are passionate about leveraging machine learning and artificial intelligence to address complex challenges in intelligence gathering, this role at ClearTrail offers an exciting opportunity to make a meaningful impact.,

You will be responsible for developing and training machine learning models that leverage drone, satellite, and real-time video data to tackle complex visual understanding tasks. Your primary focus will be on building ML/DL models for image and video classification, segmentation, and spatiotemporal analysis. Additionally, you will be tasked with developing fusion methods that align and merge data from satellite, drone, and ground footage sources. To excel in this role, you must have a minimum of 10-12 years of relevant experience. You should possess strong Python coding skills along with expertise in data manipulation using libraries such as NumPy, Pandas, and SciPy. Proficiency in vision/model libraries such as MMDetection, Detectron2, and OpenCV is essential. Moreover, you must have hands-on experience in deep learning frameworks like PyTorch and TensorFlow, particularly in developing video models with TimeSformer, SlowFast, I3D, Yolo, among others. Your expertise should extend to satellite/drone geospatial fusion, GPS/IMU metadata, and scene alignment. Familiarity with time series analysis, motion tracking, and optical flow techniques (e.g., RAFT, FlowNet) is beneficial. In addition, knowledge of remote sensing libraries like Rasterio, GDAL, GeoPandas, and xarray will be advantageous for this role. Your responsibilities will also include model evaluation and optimization across various data sources, ensuring the delivery of high-quality solutions that contribute to the future of intelligence gathering.,