Vision Intelligence Portfolio on the EDGE, on
PREM and on the CLOUD

Sensor
We have deployed solutions using 1D and 2D sensors in Visible and IR spectrum. We also have the
capability to work with 3D sensors / profilers and both hyper and multi spectral cameras.

Streaming
We are Xilinx Alliance partners and we have extensive experience in all the IPs that are part of the VCU
portfolio. We have implemented H.264 / H.265 Codecs on Zynq Ultrascale+ for large customers. We have
also developed our own IP for JPEG XS compression. We are well versed with streaming frameworks like
FFMPEG and G-Streamer.

Interface
We develop Machine Vision solutions for high speed applications in the Industrial automation space. As
part of Machine Vision product portfolio, we have developed our own CameraLink IP and Framegrabber
that can be deployed for high speed Machine Vision solutions. We have also worked with other interfaces
like MIPI – CSI and LVDS.

Algorithms
We develop standard Image processing IP blocks that run on our Framegrabber or FPGA Integrated camera
modules. We develop our own Image processing kernels or use opensource libraries like OpenCV for our
solutions. We also have extensive experience in developing, training and implementing Neural Network
models like CNN, RNN, LTSM.

FPGA Hardware
We have extensive experience in implementing H.264/H.265 Video Codec Unit (VCU) on Xilinx FPGA MPSoC
EV series and building low latency use cases for customers. We develop embedded vision solutions using
Xilinx low cost portfolio (Zynq 7 series). We also have deep expertise in VLIW, DSP kernel and application
development / validation on Versal FPGA. Our application development capability includes OpenCV,
OpenGL, OpenAMP, OpenMP and OpenCL. We are fully adept at using Xilinx tool chain like Vivado, HLS and
SDSoC.

We are focussed on applications involving high speed image processing solutions and video stream
analytics using deep learning implemented on Xilinx FPGA platforms.