A machine learning-based system for detecting leishmaniasis in microscopic images

Abstract

Background Leishmaniasis, a disease caused by a protozoan, causes numerous deaths in humans each year. After malaria, leishmaniasis is known to be the deadliest parasitic disease globally. Direct visual detection of leishmania parasite through microscopy is the frequent method for diagnosis of this disease. However, this method is time-consuming and subject to errors. This study was aimed to develop an artificial intelligence-based algorithm for automatic diagnosis of leishmaniasis. Methods We used the Viola-Jones algorithm to develop a leishmania parasite detection system. The algorithm includes three procedures: feature extraction, integral image creation, and classification. Haar-like features are used as features. An integral image was used to represent an abstract of the image that significantly speeds up the algorithm. The adaBoost technique was used to select the discriminate features and to train the classifier. Results A 65% recall and 50% precision was concluded in the detection of macrophages infected with the leishmania parasite. Also, these numbers were 52% and 71%, respectively, related to amastigotes outside of macrophages. Conclusion The developed system is accurate, fast, easy to use, and cost-effective. Therefore, artificial intelligence might be used as an alternative for the current leishmanial diagnosis methods.

Laboratory diagnostics for human Leishmania infections: a polymerase chain reaction-focussed review of detection and identification methods

“…Microscopy is of use at the genus level, but cannot be used for species differentiation as all species of Leishmania are morphologically very similar [ 74 ]. Recently, machine learning has been incorporated into microscopical examination for leishmaniasis, with sensitivity and specificity of 83% and 35%, respectively, although efficacy and speed are dependent on image quality and the particular algorithm employed [ 75 ].…”

Evaluation of the appropriate use of chest CT-Scans in the diagnosis of hospitalized patients in shiraz teaching hospitals, Southern Iran

“…besides ionizing radiation dose, requisite of contrast and their acute and long-term adverse effects, the diagnostic yield of various chest computed tomography protocol differ for different diagnosis. Among the different types of CT performed for patients in our study, the [20][21][22][23] and a single educational program do not change this behavior [24]. Defensive practice and pressure to increase emergency department turnover were supposed as a barrier to decrease CT utilization for pulmonary embolism [20] which highly apply to our crowded emergency ward.…”

“…Some measures such as higher d- dimer ordering and formal thromboembolism risk factor assessment were hypothesized to decrease inappropriate imaging utilization for pulmonary thromboembolism [ 18 , 19 ]. Studies have shown that a quality improvement program [ 20 – 23 ] and a single educational program do not change this behavior [ 24 ]. Defensive practice and pressure to increase emergency department turnover were supposed as a barrier to decrease CT utilization for pulmonary embolism [ 20 ] which highly apply to our crowded emergency ward.…”

Assessment of Deep Learning Models for Cutaneous Leishmania Parasite Diagnosis Using Microscopic Images

“…With a recall of 0.823, the model correctly detected 82.3% of the actual amastigote occurrences. Zare et al [ 31 ] created The Viola–Jones approach algorithm with an adaboost optimiser algorithm using a dataset of 300 images of positive and negative cutaneous leishmaniasis. The results showed that detecting macrophages infected with leishmania parasites had a 65% recall and 50% precision, and identifying amastigotes outside of macrophages had a 52% recall and 71% precision.…”

“…The classifier was trained using the AdaBoost algorithm after discriminative characteristics were chosen. The task of recognising amastigotes outside of macrophages had a recall of 0.520 and a precision of 0.711 [ 31 ]. The images have been captured with a smartphone at a microscopic magnification of 50 and exported in PNG format with an average resolution of 1320 px × 1900 px.…”