Implementation of Hyperspectral Technology in Skin Detection: Building Efficient Systems and Robust Models

在(zài)上(shàng)一篇文(wén)章中(zhōng)，我(wǒ)們探(tàn)討(tǎo)了高光譜成(chéng)像技術(shù)在皮(pí)膚檢(jiǎn)測中的(dí)潛(qián)力，而本文(wén)將關(guān)注如(rú)何實現(xiàn)這一技術的實現。

In the previous article, we explored the potential of hyperspectral imaging technology in skin detection. This article will focus on its practical implementation.

皮膚(fū)樣(yàng)品(pǐn)的可見(jiàn)光和近(jìn)紅(hóng)外(wài)光(guāng)譜 / Visible and Near-Infrared Spectra of Skin Samples

為(wéi)實(shí)現高光(guāng)譜(pǔ)成像(xiàng)技術的(dí)有效應用，多個(gè)研究團(tuán)隊搭建了(liǎo)各具(jù)特色的(dí)高光(guāng)譜成像係統(tǒng)。其中(zhōng)一個西(xī)班牙(yá)團(tuán)隊，搭建了不同的(dí)係統(tǒng)。他(tā)們(mén)使(shǐ)用398.08~995.20nm的高(gāo)光譜相機，配備(bèi)了(liǎo)電(diàn)動底座(zuò)和(hé)鹵素(sù)光(guāng)源，以優化(huà)成(chéng)像(xiàng)質量(liáng)，確保穩(wěn)定(dìng)的數據(jù)采集(jí)。

該團隊(duì)還(huán)搭建(jiàn)了，采(cǎi)用900~1700nm的光(guāng)譜(pǔ)範(fàn)圍，搭(dā)建係統時特別(bié)關注(zhù)患(huàn)者(zhě)的舒適度(dù)，設計(jì)了支(zhī)撐(chēng)裝置，讓患者在拍攝過程(chéng)中能夠穩定休息。此裝置(zhì)由金屬梁和(hé)多個3D打(dǎ)印支撐平台構成，提供(gōng)了柔(róu)軟且適應(yīng)不(bù)同部(bù)位(wèi)的(dí)支(zhī)持(chí)。

To effectively apply hyperspectral imaging, multiple research teams have developed specialized systems. One Spanish team, for instance, constructed distinct setups. They employed a hyperspectral camera covering 398.08–995.20 nm, equipped with a motorized stage and halogen lighting to optimize imaging quality and ensure stable data acquisition.

The team also developed another system operating in the 900–1700 nm range, prioritizing patient comfort by incorporating a support device that allowed subjects to remain stable during imaging. This setup consisted of metal beams and multiple 3D-printed support platforms, providing soft and adaptable positioning for different body areas.

可見光係(xì)統(tǒng) / the visible light system

近(jìn)紅外(wài)係統(tǒng)。（a）本研究(jiū)中(zhōng)為數據(jù)采集目的(dí)而(ér)構建(jiàn)的(dí)高光譜(pǔ)推(tuī)掃平台(tái)。（b）在采(cǎi)集過程中幫助(zhù)患者感(gǎn)到舒(shū)適的不同(tóng)支持平台(tái)。

the near-infrared system. (a) The hyperspectral push-broom platform constructed for data collection in this study. (b) Various support platforms designed to enhance patient comfort during acquisition.

在數(shù)據(jù)分(fēn)析(xī)方(fāng)法(fǎ)上，近年來(lái)的(dí)研究主要(yào)集(jí)中(zhōng)在機器學(xué)習模型(xíng)的應用。傳統(tǒng)的(dí)簡(jiǎn)單圖像(xiàng)處(chǔ)理方法雖(suī)然(rán)實(shí)現直(zhí)接(jiē)，但(dàn)在(zài)應對復(fù)雜(zá)皮(pí)膚病(bìng)變(biàn)時(shí)，其效果往(wǎng)往(wǎng)不能(néng)令人滿(mǎn)意。機(jī)器學習模型，為皮(pí)膚(fū)檢(jiǎn)測(cè)的準確(què)性提(tí)供了支(zhī)持，這(zhè)些(xiē)模型(xíng)具備(bèi)良(liáng)好的泛(fàn)化(huà)能力(lì)，能夠在多種條件(jiàn)下(xià)有(yǒu)效識別(bié)不同類型的(dí)皮膚病(bìng)變(biàn)。

Recent research has increasingly focused on machine learning models for data analysis. While traditional image processing methods are straightforward, their performance in detecting complex skin lesions is often unsatisfactory. Machine learning models, however, offer superior accuracy and generalization, enabling reliable identification of diverse skin lesions under varying conditions.

在(zài)一個研(yán)究(jiū)中，科(kē)研人員(yuán)對不(bù)同(tóng)的分(fēn)類(lèi)和分(fēn)割(gē)方(fāng)法進(jìn)行了比較。這(zhè)些(xiē)方法(fǎ)各有優缺點(diǎn)，支持(chí)向(xiàng)量機(jī)在(zài)高維(wéi)空間(jiān)中表現(xiàn)良(liáng)好，隨機森林對過擬(nǐ)合(hé)有一定的魯棒性(xìng)，K均值聚(jù)類適用(yòng)於(yú)簡單的分類任(rèn)務，而主(zhǔ)成(chéng)分分(fēn)析(xī)則(zé)有效進行降維，保(bǎo)留數(shù)據中的重要特(tè)征(zhēng)。這(zhè)具(jù)體取決於組織和(hé)目標(biāo)病(bìng)變(biàn)的(dí)類(lèi)型。

In one study, researchers compared different classification and segmentation approaches, each with unique strengths:

·Support Vector Machines (SVM) excel in high-dimensional spaces.

·Random Forests demonstrate robustness against overfitting.

·K-means Clustering is suitable for simpler classification tasks.

·Principal Component Analysis (PCA) effectively reduces dimensionality while preserving critical features.

·The optimal method depends on tissue type and the target lesion.

各類(lèi)方法(fǎ)的比(bǐ)較（部(bù)分(fēn)）/ Comparison of different methodologies (partial)

前麵(miàn)提(tí)到的西(xī)班牙團隊(duì)，該團(tuán)隊利(lì)用近紅外(wài)高(gāo)光(guāng)譜(pǔ)成像技(jì)術，針對(duì)基(jī)底細胞(bāo)癌（BCC）和皮膚鱗(lín)狀細胞(bāo)癌（SCC）進(jìn)行(háng)了檢測(cè)，強調使用(yòng)魯棒(bàng)特征統計(jì)方法(fǎ)來進行(háng)數據(jù)分析(xī)。該方法(fǎ)不僅(jǐn)提(tí)高了係統的(dí)穩定(dìng)性，還確保(bǎo)在樣本(běn)中存在噪(zào)聲和(hé)異(yì)常值(zhí)時，依舊能(néng)獲(huò)得(dé)較(jiào)高(gāo)的(dí)檢測(cè)準確性。

the aforementioned Spanish team utilized near-infrared hyperspectral imaging to detect basal cell carcinoma (BCC) and squamous cell carcinoma (SCC), emphasizing robust statistical feature extraction. This approach not only improved system stability but also maintained high detection accuracy despite noise and outliers.

上圖
：使用每個(gè)樣本(běn)在(zài)各個(gè)波長上(shàng)的(dí)中位(wèi)數值所(suǒ)得到的魯(lǔ)棒特(tè)征。

下圖
：使用(yòng)平(píng)方根的雙權(quán)重(zhòng)中(zhōng)方差作為(wéi)每個(gè)樣(yàng)本變異(yì)度的測量(liáng)方法，得到這些樣本的(dí)魯棒偏差。

Top: Robust features derived from median values of each sample across wavelengths.

Bottom: Robust deviations calculated using the square root of the biweight midvariance (√BWMV) as a measure of variability.

此外，他們在另一個(gè)實(shí)驗中重(zhòng)點關注BCC
，SCC和AK（光化性(xìng)角(jiǎo)化(huà)病）與健康皮(pí)膚的(dí)差異(yì)，同(tóng)樣(yàng)采(cǎi)用(yòng)了(liǎo)魯(lǔ)棒統計方法(fǎ)，同時還使用多變(biàn)量統計分析進行樣(yàng)本(běn)間(jiān)的比較，以發現(xiàn)數(shù)據中潛(qián)在(zài)的差(chà)異。

In another experiment, the team examined differences among BCC, SCC, actinic keratosis (AK), and healthy skin, again applying robust statistics alongside multi-variate analysis to uncover subtle data variations.

在本(běn)研(yán)究中(zhōng)通(tōng)過(guò)多(duō)種(zhǒng)方法確定的最佳(jiā)定義(yì)窗(chuāng)口(kǒu)。虛綫(xiàn)垂(chuí)直綫所(suǒ)劃定的區(qū)域標記(jì)了(liǎo)573.45nm至(zhì)779.88nm之間最終感(gǎn)興(xīng)趣(qù)的窗口。

Optimal spectral window (573.45–779.88 nm, marked by dashed vertical lines) identified through multiple methods in this study.

每個(gè)樣(yàng)本(běn)的(dí)高光(guāng)譜特征。（a）魯棒(bàng)特征標(biāo)記了(liǎo)中央(yāng)傾(qīng)向(xiàng)以及5%和(hé)95%百分(fēn)位(wèi)置信區(qū)間（下綫(xiàn)和上綫分(fēn)別）。（b）√BWMV計算(suàn)表(biǎo)示(shì)魯棒樣本(běn)方(fāng)差(chà)。

Hyperspectral features of each sample. (a) Robust features indicating central tendency with 5% and 95% percentile confidence intervals (lower and upper bounds, respectively). (b) √BWMV representing robust sample variance.

綜(zōng)上所述(shù)，高(gāo)光譜(pǔ)成像技術在皮(pí)膚檢測中(zhōng)展(zhǎn)現出了優(yōu)勢，尤其是在(zài)係統構建(jiàn)與模型泛(fàn)化能(néng)力(lì)方麵。通過選(xuǎn)擇(zé)適宜(yí)的(dí)波(bō)長(cháng)範圍，結合(hé)先(xiān)進的數據分析(xī)技術，我們的高光譜相(xiāng)機在皮(pí)膚疾病早(zǎo)期檢(jiǎn)測(cè)中提(tí)供了堅實(shí)的(dí)基礎(chǔ)。

值得(dé)一(yī)提的(dí)是，我們公(gōng)司(sī)不僅銷售(shòu)高(gāo)光譜相(xiāng)機(jī)，還(huán)能提供專(zhuān)業的(dí)硬件技(jì)術支(zhī)持，助(zhù)力您的(dí)研(yán)究(jiū)與(yǔ)應(yīng)用提(tí)升效(xiào)率。未來，隨(suí)著(zhuó)高(gāo)光譜成像(xiàng)技術與機(jī)器學(xué)習的深(shēn)度融(róng)合，該領域必將(jiāng)迎(yíng)來(lái)更(gēng)多機會(huì)，相(xiāng)信皮(pí)膚(fū)癌的早期檢測將(jiāng)變得(dé)更(gēng)加高(gāo)效和可靠，為患(huàn)者帶(dài)來更大的福音(yīn)。

Hyperspectral imaging demonstrates unique advantages in skin detection, particularly in system design and model generalization. By selecting optimal wavelength ranges and integrating advanced analytics, hyperspectral cameras provide a robust foundation for early skin disease diagnosis.

Notably, our company not only supplies hyperspectral cameras but also offers expert hardware support to enhance research and application efficiency. As hyperspectral imaging and machine learning continue to converge, this field holds immense promise—ushering in more efficient, reliable early detection of skin cancer and greater benefits for patients.

案(àn)例來源 / Source
：

1. Courtenay LA, González-Aguilera D, Lagüela S, Del Pozo S, Ruiz-Mendez C, Barbero-García I, Román-Curto C, Cañueto J, Santos-Durán C, Cardeñoso-Álvarez ME, Roncero-Riesco M, Hernandez-Lopez D, Guerrero-Sevilla D, Rodríguez-Gonzalvez P. Hyperspectral imaging and robust statistics in non-melanoma skin cancer analysis. Biomed Opt Express. 2021 Jul 20;12(8):5107-27. doi: 10.1364/BOE.428143. PMID: 34513245; PMCID: PMC8407807.

2. Courtenay LA, Barbero-García I, Martínez-Lastras S, Del Pozo S, Corral de la Calle M, Garrido A, Guerrero-Sevilla D, Hernandez-Lopez D, González-Aguilera D. Near-infrared hyperspectral imaging and robust statistics for in vivo non-melanoma skin cancer and actinic keratosis characterisation. PLoS One. 2024 Apr 25;19(4):e0300400. doi: 10.1371/journal.pone.0300400. PMID: 38662718; PMCID: PMC11045066.

3. Aloupogianni E, Ishikawa M, Kobayashi N, Obi T. Hyperspectral and multispectral image processing for gross-level tumor detection in skin lesions: a systematic review. J Biomed Opt. 2022 Jun 8;27(6):060901. doi: 10.1117/1.JBO.27.6.060901.