Hyper spectral images (HSI) provide rich spectral and spatial information across a series of contiguous spectral bands. However, the accurate processing of the spectral and spatial correlation between the bands requires the use of energy-expensive 3-D Convolutional Neural Networks (CNNs). To address this challenge, we propose the use of Spiking Neural Networks (SNNs) that are generated from iso-architecture CNNs and trained with quantization-aware gradient descent to optimize their weights, membrane leak, and firing thresholds. During both training and inference, the analog pixel values of a HSI are directly applied to the input layer of the SNN without the need to convert to a spike-train. The reduced latency of our training technique combined with high activation sparsity yields significant improvements in computational efficiency. We evaluate our proposal using three HSI datasets on a 3-D and a 3-D / 2-D hybrid convolutional architecture. We achieve overall accuracy, average accuracy, and kappa coefficient of 98.68%, 98.34%, and 98.20% respectively with 5 time steps (inference latency) and 6-bit weight quantization on the Indian Pines dataset. In particular, our models achieved accuracies similar to state-of-the-art (SOTA) with 560.6 and 44.8 times less compute energy on average over three HSI datasets than an iso-architecture full-precision and 6-bit quantized CNN, respectively.

中文翻译：

---

面向高光谱图像分类的节能量化深度尖峰神经网络

超光谱图像 (HSI) 提供了一系列连续光谱带的丰富光谱和空间信息。然而，波段之间光谱和空间相关性的准确处理需要使用能量昂贵的 3-D 卷积神经网络 (CNN)。为了应对这一挑战，我们建议使用由 iso-architecture CNN 生成并经过量化感知梯度下降训练的尖峰神经网络 (SNN) 来优化它们的权重、膜泄漏和触发阈值。在训练和推理期间，HSI 的模拟像素值直接应用于 SNN 的输入层，无需转换为尖峰训练。我们的训练技术的延迟降低与高激活稀疏性相结合，显着提高了计算效率。我们在 3-D 和 3-D / 2-D 混合卷积架构上使用三个 HSI 数据集评估我们的建议。我们在印度松树数据集上通过 5 个时间步长（推理延迟）和 6 位权重量化分别实现了 98.68%、98.34% 和 98.20% 的整体准确度、平均准确度和 kappa 系数。特别是，我们的模型在三个 HSI 数据集上的平均计算能量分别比 iso-architecture 全精度和 6 位量化 CNN 低 560.6 和 44.8 倍，达到了与最先进技术 (SOTA) 相似的精度。印度松树数据集上的 5 个时间步长（推理延迟）和 6 位权重量化分别为 20%。特别是，我们的模型在三个 HSI 数据集上的平均计算能量分别比 iso-architecture 全精度和 6 位量化 CNN 低 560.6 和 44.8 倍，达到了与最先进技术 (SOTA) 相似的精度。印度松树数据集上的 5 个时间步长（推理延迟）和 6 位权重量化分别为 20%。特别是，我们的模型在三个 HSI 数据集上的平均计算能量分别比 iso-architecture 全精度和 6 位量化 CNN 低 560.6 和 44.8 倍，达到了与最先进技术 (SOTA) 相似的精度。