Classifying microscopic images as acute lymphoblastic leukemia by Resnet ensemble model and Taguchi method

Background Researchers have attempted to apply deep learning methods of artificial intelligence for rapidly and accurately detecting acute lymphoblastic leukemia (ALL) in microscopic images. Results A Resnet101-9 ensemble model was developed for classifying ALL in microscopic images. The proposed Resnet101-9 ensemble model combined the use of the nine trained Resnet-101 models with a majority voting strategy. Each trained Resnet-101 model integrated the well-known pre-trained Resnet-101 model and its algorithm hyperparameters by using transfer learning method to classify ALL in microscopic images. The best combination of algorithm hyperparameters for the pre-trained Resnet-101 model was determined by Taguchi experimental method. The microscopic images used for training of the pre-trained Resnet-101 model and for performance tests of the trained Resnet-101 model were obtained from the C-NMC dataset. In experimental tests of performance, the Resnet101-9 ensemble model achieved an accuracy of 85.11% and an F1-score of 88.94 in classifying ALL in microscopic images. The accuracy of the Resnet101-9 ensemble model was superior to that of the nine trained Resnet-101 individual models. All other performance measures (i.e., precision, recall, and specificity) for the Resnet101-9 ensemble model exceeded those for the nine trained Resnet-101 individual models. Conclusion Compared to the nine trained Resnet-101 individual models, the Resnet101-9 ensemble model had superior accuracy in classifying ALL in microscopic images obtained from the C-NMC dataset.

CFD simulations of rotary BOF slag carbonation kiln reactor with cyclone flow

Rotary kiln reactors play an important role in improving the mechanical properties and usability of basic oxygen furnace slag through a carbonation process. The performance of such reactors is critically dependent on the residence time of the CO2 gas used to promote the carbonation reaction. Accordingly, the present study proposes a rotary kiln reactor in which the residence time is increased by arranging the inlet and outlet pipes obliquely to the reactor centerline; thereby producing a cyclone flow structure within the reactor tube. The optimal geometry parameters and rotational speed of the kiln are determined using the robust Taguchi experimental method. The CO2 residence time in the optimized kiln is then evaluated by means of computational fluid dynamics simulations. It is shown that the residence time increases from 63.587 s in a standard (non-cyclone-flow) rotary kiln to 105.815 s in the optimized rotary kiln; corresponding to a performance improvement of 66.4%.

Impact of Laser Parameters on the Contrasts of Laser Direct Marked Data Matrix Symbols on Titanium Alloy Substrates

To determine the impact of laser parameters on the contrasts of laser direct marked Data Matrix symbols on titanium alloys, a Q-switched Nd:YAG laser was used in the marking process. For this purpose, four laser marking parameters (i.e. electric current, effective vector step, Q-switch frequency, and laser line spacing) were correlated with the symbol contrast (SC). The L25 orthogonal array based on the Taguchi experimental method was adopted to determine the optimal combination levels of laser parameters for the SC, and the experimental data were statistically analyzed by multi-factor analysis of variance (ANOVA). Experimental results showed that the electric current, effectie vector step, laser line spacing have a statistically significant impact on the contrasts of laser marked Data Matrix symbols. Q-switch frequency is statistically insignificant at a 5% level. The optimal combination levels of laser parameters for the SC is where the electric current is at 26A, the effective vector step is at 0.001mm, the laser line spacing is at 0.01mm, and Q-switch frequency is at 5kHz.