[论文解读] Disease Classification within Dermascopic Images Using features extracted by ResNet50 and classification through Deep Forest

Sounak Ray|arXiv (Cornell University)|Jul 16, 2018

Cutaneous Melanoma Detection and Management被引用 4

一句话总结

本文提出了一种混合深度学习方法，用于通过预训练的ResNet50进行特征提取，并利用基于树的集成模型Deep Forest进行分类，实现皮肤病变分类。该方法在ISIC 2018数据集上表现出色，尤其在小样本和数据不平衡场景下，得益于Deep Forest的自适应复杂度和对类别不平衡的鲁棒性，取得了具有竞争力的性能。

ABSTRACT

In this report we propose a classification technique for skin lesion images as a part of our submission for ISIC 2018 Challenge in Skin Lesion Analysis Towards Melanoma Detection. Our data was extracted from the ISIC 2018: Skin Lesion Analysis Towards Melanoma Detection grand challenge datasets. The features are extracted through a Convolutional Neural Network, in our case ResNet50 and then using these features we train a DeepForest, having cascading layers, to classify our skin lesion images. We know that Convolutional Neural Networks are a state-of-the-art technique in representation learning for images, with the convolutional filters learning to detect features from images through backpropagation. These features are then usually fed to a classifier like a softmax layer or other such classifiers for classification tasks. In our case we do not use the traditional backpropagation method and train a softmax layer for classification. Instead, we use Deep Forest, a novel decision tree ensemble approach with performance highly competitive to deep neural networks in a broad range of tasks. Thus we use a ResNet50 to extract the features from skin lesion images and then use the Deep Forest to classify these images. This method has been used because Deep Forest has been found to be hugely efficient in areas where there are only small-scale training data available. Also as the Deep Forest network decides its complexity by itself, it also caters to the problem of dataset imbalance we faced in this problem.

研究动机与目标

• 解决在训练数据有限的情况下，对皮肤镜图像中的皮肤病变进行分类的挑战。
• 克服传统深度学习分类器（如Softmax）在小样本和数据不平衡数据集中的局限性。
• 探索Deep Forest作为ResNet50提取特征的替代分类方法的有效性。
• 通过结合迁移学习与一种稳健且自适应的集成模型，提升黑色素瘤检测的准确率。

提出的方法

• 使用预训练的ResNet50模型从皮肤镜图像中提取特征。
• 将ResNet50最后一个全连接层学习到的特征用作分类输入。
• 在这些提取的特征上训练Deep Forest模型，而非使用传统的Softmax分类器。
• Deep Forest通过级联多层随机森林学习分层表示并进行预测。
• 该模型可根据数据特征自动调整其复杂度，从而在小样本数据集中减少过拟合。
• 该方法避免了分类过程中的反向传播，转而依赖基于树的集成学习，从而提升鲁棒性。

实验结果

研究问题

• RQ1 当与预训练ResNet50提取的特征结合使用时，Deep Forest是否能优于传统的Softmax分类器？
• RQ2 ResNet50特征提取与Deep Forest分类相结合，在处理小样本皮肤镜图像数据集方面有多高效？
• RQ3 Deep Forest在多大程度上缓解了皮肤病变数据集中常见的类别不平衡问题？
• RQ4 Deep Forest的自适应复杂度是否能提升在有限训练数据上的泛化能力？

主要发现

• 所提出的方法在ISIC 2018数据集上实现了具有竞争力的分类性能，表现出在数据有限情况下的强泛化能力。
• Deep Forest能有效处理数据集不平衡问题，减少对多数类别的偏差。
• 模型的自动复杂度调节机制可防止过拟合，尤其在小样本训练设置中优势明显。
• 与在相同特征上应用的标准Softmax分类相比，该混合方法在低数据场景下表现更优。