High Efficiency Video Coding (HEVC) provides more compression than its predecessors. One of the modules that contributes to higher compression rates is the Motion Estimation module, which consists of Integer and Fractional pixel motion estimation. The Fractional Motion Estimation (FME) process performs interpolations to find sample values at fractional-pixel locations, which can be computationally demanding. In this paper, we propose an interpolation-free method for FME based on Artificial Neural Networks (ANNs). Our proposed method is implemented in HEVC reference software (HM-16.9). According to our results, ANNs can accomplish FME task with an average increase of 2.6% in BDRate and an average reduction of 0.09 dB in BD-PSNR.

Efficient implementation of non-linear activation functions is essential to the implementation of deep learning models on FPGAs. We introduce such an implementation based on the Discrete Cosine Transform Interpolation Filter (DCTIF). The proposed interpolation architecture combines simple arithmetic operations on the stored samples of the hyperbolic tangent function and on input data. It achieves almost 3× better precision than previous works while using a similar amount computational resources and a small amount of memory. Various combinations of DCTIF parameters can be chosen to trade off the accuracy and the overall circuit complexity of the tanh function. In one case, the proposed architecture approximates the hyperbolic tangent activation function with 0.004 maximum error while requiring only 1.45 kbits BRAM memory and 21 LUTs of a Virtex-7 FPGA.