Dilate-Invariant Temporal Convolutional Network for Real-Time Edge Applications

Emad Ibrahim (Corresponding author), Bart van den Dool, Sayandip De, Manil Dev Gomony, Jos A. Huisken, Marc C.W. Geilen

Research output: Contribution to journalArticleAcademicpeer-review

25 Downloads (Pure)


Temporal Convolutional Networks (TCNs) involving mono channels as input, have shown superior performance compared to state-of-the-art sequence detection recursive networks in a variety of applications. TCNs leverage the concept of dilated causal convolution for a wider receptive field coverage of input (mono) channels, which requires scaling the delay between input samples in Multiply-Accumulate (MAC) units in different layers. We demonstrate a possible data-flow transformation to convert a dilated convolution to a non-dilated convolution to remove such need for delay scaling while maintaining the same receptive field. The new data-flow transformation allows for hardware units to be shared across all layers with single-delay units between the MAC units. We demonstrate how such data-flow transformation can be easily achieved using generic Finite Impulse Response (FIR) filter modules, simplifying the deployment of TCNs. We validate the predicted savings using Cadence Stratus High-Level Synthesis (HLS). A gesture recognition case study using ultrasound is synthesized achieving 25% savings in both energy and area if the data-flow transformation is applied.
Original languageEnglish
Pages (from-to)1210-1220
Number of pages11
JournalIEEE Transactions on Circuits and Systems I: Regular Papers
Issue number3
Publication statusPublished - 1 Mar 2022


  • Temporal convolutional network
  • data-flow transformations
  • digital signal processing
  • dilated convolution
  • high-level synthesis


Dive into the research topics of 'Dilate-Invariant Temporal Convolutional Network for Real-Time Edge Applications'. Together they form a unique fingerprint.

Cite this