Burn characterization using object-oriented hyperspectral image classification

A systematic review of advances in the use of spectral imaging in burn depth assessment

BACKGROUND

Accurate burn depth assessment is critical for determining appropriate treatment and optimizing patient outcomes. Conventional methods, such as clinical assessment and laser Doppler imaging, have limitations in terms of accuracy and timeliness. Spectral imaging, including multispectral imaging and hyperspectral imaging, has emerged as a promising non-invasive modality to improve burn depth evaluation. This systematic review aims to evaluate the advances in spectral imaging technologies for burn depth assessment, with a focus on diagnostic accuracy, the role of machine learning integration, and the quality of current evidence.

METHODS

A comprehensive literature search was conducted in March 2024 using PubMed, Scottish Network, EMBASE, and Cochrane Library databases. Studies that evaluated spectral imaging for burn depth assessment and compared it to standard methods such as laser Doppler imaging, clinical assessment, or histological analysis were included. The quality of the included studies was assessed using the QUADAS tool.

RESULTS

Seven studies from 1988 to 2023 met the inclusion criteria, evaluating a total of 167 patients with 269 burn sites. The pooled analysis revealed a combined sensitivity of 86 % (95 % CI [0.80; 0.90]) and specificity of 84 % (95 % CI [0.70; 0.93]. However, there was a large range of sensitivity identified from 61 % to 97.2 % and specificity from 45 % to 100 %. Notably, the integration of machine learning, particularly convolutional neural networks and support vector machines, improved classification accuracy, with some models achieving over 95 % sensitivity and specificity. Despite these promising results, significant variability in methodologies and a lack of standardized ground truthing were identified.

CONCLUSION

Spectral imaging, especially when combined with machine learning, shows strong potential as an effective tool for burn depth assessment, offering high diagnostic accuracy and reproducibility. Further research is needed to standardize protocols and validate these technologies across diverse patient populations, paving the way for clinical adoption and improved patient care.

Automatic detection of ischemic necrotic sites in small intestinal tissue using hyperspectral imaging and transfer learning

Acquiring large amounts of hyperspectral data of small intestinal tissue with real labels in the clinic is difficult, and the data shows inter-patient variability. Building an automatic identification model using a small dataset presents a crucial challenge in obtaining a strong generalization of the model. This study aimed to explore the performance of hyperspectral imaging and transfer learning techniques in the automatic identification of normal and ischemic necrotic sites in small intestinal tissue. Hyperspectral data of small intestinal tissues were collected from eight white rabbit samples. The transfer component analysis (TCA) method was performed to transfer learning on hyperspectral data between different samples and the variability of data distribution between samples was reduced. The results showed that the TCA transfer learning method improved the accuracy of the classification model with less training data. This study provided a reliable method for single-sample modelling to detect necrotic sites in small intestinal tissue .

Discrimination between normal and necrotic small intestinal tissue using hyperspectral imaging and unsupervised classification

Objective and automatic clinical discrimination of normal and necrotic sites of small intestinal tissue remains challenging. In this study, hyperspectral imaging (HSI) and unsupervised classification techniques were used to distinguish normal and necrotic sites of small intestinal tissues. Small intestinal tissue hyperspectral images of eight Japanese large-eared white rabbits were acquired using a visible near-infrared hyperspectral camera, and K-means and density peaks (DP) clustering algorithms were used to differentiate between normal and necrotic tissue. The three cases in this study showed that the average clustering purity of the DP clustering algorithm reached 92.07% when the two band combinations of 500-622 and 700-858 nm were selected. The results of this study suggest that HSI and DP clustering can assist physicians in distinguishing between normal and necrotic sites in the small intestine in vivo.