​InnoWave has unmatched expertise in the world of vision, in-depth expertise in the areas of computer vision and firmware development

for video processing, embedded systems and many others.

 

InnoWave offers both R&D and fully functional computer vision software development services. We use CV libraries such as OpenCV, optimize existing methods, and create custom algorithms based on mathematical models. We also apply AI techniques, deep learning, and machine learning. We are experienced in GPUs accelerating software, optimizing neural network interface performance through TensorRT for a variety of machine learning frameworks, including TensorFlow, PyTrorch, Caffe, and others. 

We build and train models using supervised (classification or regression-based), unsupervised (clustering or association based) and reinforcement (reaction to an environment based) learning techniques.

We’ve a wide experience with Nvidia Jetson platforms, Intel, Qualcomm, AMD/Xilinx AI Edge, NXP, MediaTek and many more.

​

​

​​

​

​​​​​​

• Custom Neural Network development.

• Running inference in realtime - object detection, object tracking, human pose estimation.

• Model benchmarking & optimisations.

•  Realtime video encoding/decoding.

• Building complex Deepstream pipelines.

• Realtime processing of multiple cameras.

• Working with hardware-synced cameras

• Running computer vision tasks in realtime - depth estimation, optical flow, etc

• BSP customization, driver support for cameras and other peripherals, like gyro, accelerometers, GPS, GMSL links.

 

​