ImageHD: Energy-Efficient On-Device Continual Learning of Visual Representations via Hyperdimensional Computing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Arockiaraj, Jebacyril, Parikh, Dhruv, Prasanna, Viktor
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914501011439616
author Arockiaraj, Jebacyril
Parikh, Dhruv
Prasanna, Viktor
author_facet Arockiaraj, Jebacyril
Parikh, Dhruv
Prasanna, Viktor
contents On-device continual learning (CL) is critical for edge AI systems operating on non-stationary data streams, but most existing methods rely on backpropagation or exemplar-heavy classifiers, incurring substantial compute, memory, and latency overheads. Hyperdimensional computing (HDC) offers a lightweight alternative through fast, non-iterative online updates. Combined with a compact convolutional neural network (CNN) feature extractor, HDC enables efficient on-device adaptation with strong visual representations. However, prior HDC-based CL systems often depend on multi-tier memory hierarchies and complex cluster management, limiting deployability on resource-constrained hardware. We present ImageHD, an FPGA accelerator for on-device continual learning of visual data based on HDC. ImageHD targets streaming CL under strict latency and on-chip memory constraints, avoiding costly iterative optimization. At the algorithmic level, we introduce a hardware-aware CL method that bounds class exemplars through a unified exemplar memory and a hardware-efficient cluster merging strategy, while incorporating a quantized CNN front-end to reduce deployment overhead without sacrificing accuracy. At the system level, ImageHD is implemented as a streaming dataflow architecture on the AMD Zynq ZCU104 FPGA, integrating HDC encoding, similarity search, and bounded cluster management using word-packed binary hypervectors for massively parallel bitwise computation within tight on-chip resource budgets. On CORe50, ImageHD achieves up to 40.4x (4.84x) speedup and 383x (105.1x) energy efficiency over optimized CPU (GPU) baselines, demonstrating the practicality of HDC-enabled continual learning for real-time edge AI.
format Preprint
id arxiv_https___arxiv_org_abs_2604_21280
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ImageHD: Energy-Efficient On-Device Continual Learning of Visual Representations via Hyperdimensional Computing
Arockiaraj, Jebacyril
Parikh, Dhruv
Prasanna, Viktor
Computer Vision and Pattern Recognition
On-device continual learning (CL) is critical for edge AI systems operating on non-stationary data streams, but most existing methods rely on backpropagation or exemplar-heavy classifiers, incurring substantial compute, memory, and latency overheads. Hyperdimensional computing (HDC) offers a lightweight alternative through fast, non-iterative online updates. Combined with a compact convolutional neural network (CNN) feature extractor, HDC enables efficient on-device adaptation with strong visual representations. However, prior HDC-based CL systems often depend on multi-tier memory hierarchies and complex cluster management, limiting deployability on resource-constrained hardware. We present ImageHD, an FPGA accelerator for on-device continual learning of visual data based on HDC. ImageHD targets streaming CL under strict latency and on-chip memory constraints, avoiding costly iterative optimization. At the algorithmic level, we introduce a hardware-aware CL method that bounds class exemplars through a unified exemplar memory and a hardware-efficient cluster merging strategy, while incorporating a quantized CNN front-end to reduce deployment overhead without sacrificing accuracy. At the system level, ImageHD is implemented as a streaming dataflow architecture on the AMD Zynq ZCU104 FPGA, integrating HDC encoding, similarity search, and bounded cluster management using word-packed binary hypervectors for massively parallel bitwise computation within tight on-chip resource budgets. On CORe50, ImageHD achieves up to 40.4x (4.84x) speedup and 383x (105.1x) energy efficiency over optimized CPU (GPU) baselines, demonstrating the practicality of HDC-enabled continual learning for real-time edge AI.
title ImageHD: Energy-Efficient On-Device Continual Learning of Visual Representations via Hyperdimensional Computing
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.21280