The Transformer-based target detection model, DETR, has powerful feature extraction and recognition capabilities, but its high computational and storage requirements limit its deployment on resource-constrained devices. To solve this problem, we first replace the ResNet-50 backbone network in DETR with Swin-T, which realizes the unification of the backbone network with the Transformer encoder and decoder under the same Transformer processing paradigm. On this basis, we propose a quantized inference scheme based entirely on integers, which effectively serves as a data compression method for reducing memory occupation and computational complexity. Unlike previous approaches that only quantize the linear layer of DETR, we further apply integer approximation to all non-linear operational layers (e.g., Sigmoid, Softmax, LayerNorm, GELU), thus realizing the execution of the entire inference process in the integer domain. Experimental results show that our method reduces the computation and storage to 6.3% and 25% of the original model, respectively, while the average accuracy decreases by only 1.1%, which validates the effectiveness of the method as an efficient and hardware-friendly solution for target detection.

The rise of the Internet of Things (IoT) and edge computing increases the demand for efficient image transmission under strict latency, memory, and power budgets. While learned semantic communication can reduce bandwidth by transmitting task-relevant representations rather than raw pixels, state-of-the-art models are difficult to deploy on resource-constrained devices due to their reliance on floating-point computation and heavy memory footprint. To bridge this gap, we propose a fully quantized lightweight image semantic communication system (FQ-LISCS), which constructs a full-path integer-only inference pipeline for semantic image transmission. Specifically, FQ-LISCS applies 8-bit integer (INT8) quantization to linear operators and introduces efficient integer approximations for key nonlinearities, enabling end-to-end deployment without floating-point operations during inference. Experimental results demonstrate that FQ-LISCS preserves transmission quality close to full-precision baselines, with a peak signal-to-noise ratio (PSNR) drop of only 0.18 dB compared with wireless image transmission transformer (WITT) on the Kodak test set, while substantially reducing hardware resource consumption.