Transfer Learning-Based Condition Monitoring of Single Point Cutting Tool.

Machining activities in recent times have shifted their focus towards tool life and tool wear. Cutting tools have been utilized on a daily basis and play a vital role in manufacturing industries. Prolonged and incessant operation of the cutting tool can lead to wear and tear of the component, thereby compromising the dimensional accuracy. The condition of a tool is estimated based upon the surface quality of the machined component, condition of the machine, and the rate of production. Maintaining the tool health plays a vital role in enhancing the productivity of manufacturing industries. Numerous efforts were experimented by the researchers to maintain the tool health condition. The drawbacks of conventional diagnostic techniques include requirement of high level of human intelligence and professional expertise on the field, which led the researchers to develop intelligent and automatic diagnostic tools. There are many techniques suggested by researchers to detect the condition of single point cutting tool. This article proposes the use of transfer learning technology to detect the condition of single point cutting tool. First, the vibration signals were collected from the cutting tool and plots were made which will work as input to the deep learning algorithms. The deep learning algorithms have the capability to learn from the plots of vibration signals and classify the state of the single point cutting tool. In this work, the pretrained networks such as VGG-16, AlexNet, ResNet-50, and GoogLeNet were employed to identify the state of the cutting tool. In the pretrained networks, the effect of hyperparameters such as batch size, solver, learning rate, and train-test split ratio was studied, and the best performing network was suggested for tool condition monitoring.

Misfire Detection in Spark Ignition Engine Using Transfer Learning.

Misfire detection in an internal combustion engine is an important activity. Any undetected misfire can lead to loss of fuel and power in the automobile. As the fuel cost is more, one cannot afford to waste money because of the misfire. Even if one is ready to spend more money on fuel, the power of the engine comes down; thereby, the vehicle performance falls drastically because of the misfire in IC engines. Hence, researchers paid a lot of attention to detect the misfire in IC engines and rectify it. Drawbacks of conventional diagnostic techniques include the requirement of high level of human intelligence and professional expertise in the field, which made the researchers look for intelligent and automatic diagnostic tools. There are many techniques suggested by researchers to detect the misfire in IC engines. This paper proposes the use of transfer learning technology to detect the misfire in the IC engine. First, the vibration signals were collected from the engine head and plots are made which will work as input to the deep learning algorithms. The deep learning algorithms have the capability to learn from the plots of vibration signals and classify the state of the misfire in the IC engines. In the present work, the pretrained networks such as AlexNet, VGG-16, GoogLeNet, and ResNet-50 are employed to identify the misfire state of the engine. In the pretrained networks, the effect of hyperparameters such as back size, solver, learning rate, and train-test split ratio was studied and the best performing network was suggested for misfire detection.

Fluorescence spectra of blood components for breast cancer diagnosis.

OBJECTIVE: To explore whether fluorescence emission spectroscopy of blood components can differentiate normal from early and advanced stages of breast cancer using stepwise discriminant analysis. MATERIALS AND METHODS: Fluorescence emission spectra were measured for blood components of three different groups: 35 normal controls, 28 with early-stage, and 18 with advanced-stage breast cancer. The data from the spectra were subjected to Fisher's linear discriminant analysis. Classification accuracy, specificity, and sensitivity of the technique were calculated for breast cancer diagnosis. RESULTS: Fluorescence emission spectra of blood components accurately distinguished normal from early-stage and advanced-stage breast cancer in 91.4% of original cases and 90.1% for cross-validated cases. The sensitivity and specificity were 80.4% and 100%, respectively, in distinguishing subjects with breast cancer from normal controls. CONCLUSION: Our statistical evaluation indicates that porphyrin in blood can be used as a reliable tumor marker. Fluorescence emission spectroscopy of blood components and statistical evaluations should be further investigated for a variety of tumors.

Detection of cancer by optical analysis of body fluids--a single blind study.

This paper pertains to a new technique based on fluorescence emission spectra (FES), and stokes shift spectra (SSS) of blood plasma, acetone extract of cellular fraction, and urine. These samples were collected from 60 cancer patients of different etiology and 60 age adjusted controls for a single blind study. A set of ratio parameters were obtained from the above spectra (FES and SSS of above three sets of samples), based on the relative intensity of biofluorophores like tryptophan, tyrosine, flavin etc. It was found that these biofluorophores go out of proportion for malignancy of any etiolology. The study was done in two phases: calibration and validation. Based on a certain set of ratios obtained by simple statistical analysis, in the calibration phase, the blinded samples of validation phase were spectrally analysed and classified as normal or malignant. The scoring done by independent oncologists (who were not involved in any part of this new technique) yielded an overall sensitivity of 87%, and specificity of 83%. The result indicate that new optical spectroscopic techniques could be a simple, non-invasive protocol for detection of cancers, particularly in symptomatic cases; or for monitoring the post treated cases of cancer.

Nanosecond fluorescence resonance energy transfer-fluorescence lifetime imaging microscopy to localize the protein interactions in a single living cell.

Visualizing and quantifying protein-protein interactions is a recent trend in biomedical imaging. The current advances in fluorescence microscopy, coupled with the development of new fluorescent probes such as green fluorescent proteins, allow fluorescence resonance energy transfer (FRET) to be used to study protein interactions in living specimens. Intensity-based FRET microscopy is limited by spectral bleed-through and fluorophore concentration. Fluorescence lifetime imaging (FLIM) microscopy and lifetime measurements are independent of change in fluorophore concentration or excitation intensity, and the combination of FRET and FLIM provides high spatial (nanometre) and temporal (nanoseconds) resolution. Because only the donor fluorophore lifetime is measured, spectral bleed-through is not an issue in FRET-FLIM imaging. In this paper we describe the development of a nanosecond FRET-FLIM microscopy instrumentation to acquire the time-resolved images of donor in the presence and the absence of the acceptor. Software was developed to process the acquired images for single and double exponential decays. Measurement of donor lifetime in two different conditions allowed us to calculate accurately the distance between the interacting proteins. We used this approach to quantify the dimerization of the transcription factor CAATT/enhancer binding protein alpha in living pituitary cells. The one- and two-component analysis of the donor molecule lifetime in the presence of acceptor demonstrates the distance distribution between interacting proteins.