WAY DIFFERENTIAL DIAGNOSIS follicular adenoma and follicular thyroid carcinoma

摘要

a differential diagnosis of cancer and u0444u043eu043bu043bu0438u043au0443u043bu00a0u0440u043du043eu0433u043e u0444u043eu043bu043bu0438u043au0443u043bu00a0u0440u043du043eu0439 adenoma thyroid tumors, including thyroid u043fu0430u0446u0438u0435u043d fence material the way u0430u0441u043fu0438u0440u0430u0446u0438u043eu043du043du043eu0439 u0442u043eu043du043au043eu0438u0433u043eu043bu044cu043du043eu0439 lp and / or u0441u043eu0441u043au043eu0431u0430 with incision tissue tumor in u0432u0440u0435u043cu00a0 remote operation; making strokes, tumorstheir colouring u0434u043bu00a0 u0446u0438u0442u043eu043bu043eu0433u0438u0447u0435u0441u043au043eu0433u043e analysis; obtaining digital images of strokes, tumors using system analysis u0438u0437u043eu0431u0440u0430u0436u0435u043du0438u00a0, u0441u043eu0441u0442u043eu00a0u0449u0435u0439 of u043fu0440u043eu0445u043eu0434u00a0u0449u0435u0433 microscope o light, digital cameras, computer program and image analysis; the analysis of digital images of the tumor with the help of computer programs u0434u0432u0443u0445u0441u043bu043eu0439u043du043eu0439 strokes we like the "neural network".u0444u043eu043bu043bu0438u043au0443u043bu00a0u0440u043du044bu0439 pre trained to distinguish between cancer and u0444u043eu043bu043bu0438u043au0443u043bu00a0u0440u043du0443u044e an adenoma of thyroid, based on the reference image with the u0433u0438u0441u0442u043eu043bu043eu0433u0438u0447u0435u0441u043au0438 strokes creation u043eu0442u043bu0438u0447u0430u044eu0449u0438u0439u0441u00a0 diagnosis, so that the source digital u0438u0437u043eu0431u0440u0430u0436u0435u043du0438u00a0 strokes, tumors are growing under the microscope at 400 times to permit 13881040 u043fu0438u043au0441u0435u043bu0435 third,the range of spatial frequencies of 1 to 128 in relative units, reduce to the size of the u0438u0437u043eu0431u0440u0430u0436u0435u043du0438u00a0 256256 pixels; u043fu0440u043eu0432u043eu0434u00a0u0442 automatic the diagnosis of tumors in the u0438u0437u043eu0431u0440u0430u0436u0435u043du0438u00a0u043c painted strokes through a computer program, including u043fu0440u0435u0434u043eu0431u0440u0430u0431u043eu0442u0447u0438u043a,automatically on the basis of the analysis of u0432u044bu0434u0435u043bu00a0u044eu0449u0438u0439 fourier transform source image the most significant sign of them in the range of spatial frequencies of 96 to 128 u0443u0441u043bu043eu0432u043d s units, u043fu043eu0437u0432u043eu043bu00a0u044eu0449u0438u0435 share these u0438u0437u043eu0431u0440u0430u0436u0435u043du0438u00a0 two classthe u0444u043eu043bu043bu0438u043au0443u043bu00a0u0440u043du043eu0439 u0430u0434u0435u043du043eu043cu0435 or u0444u043eu043bu043bu0438u043au0443u043bu00a0u0440u043du043eu043cu0443 cancer; with the help of a computer program to each reference image in accordance u0441u0442u0430u0432u00a0u0442 training sample u0432u0435u0442u0441u0442u0432u0438u0435 neuron first u0441u043bu043eu00a0; space converted images of neurons first u0441u043bu043eu00a0 value e u0440u0430u0441u0441u0442u043eu00a0u043du0438u0435 from each reference u0438u0437u043eu0431u0440u0430u0436u0435u043du0438u00a0 the training sample to test u0438u0437u043eu0431u0440u0430u0436u0435u043du0438u00a0,the obtained estimates will a positive or a negative sign, depending on the class (type of tumor).which is the reference image; among the first u0441u043bu043eu00a0 neurons in each of the two classes u0432u044bu00a0u0432u043bu00a0u044eu0442 several winners of the minimum u044du0432u043au043bu0438u0434u043eu0432u0430 u0440u0430u0441u0441u0442u043eu00a0u043du0438u00a0 from the reference u0438u0437u043eu0431u0440u0430u0436u0435u043du0438u00a0 prior to testing, using only the second u0441u043bu043eu00a0 neuron u0441u0443u043cu043cu0438u0440u0443u044eu0442 inverse values with the respective u044du0432u043au043bu0438u0434u043eu0432u044bu0445 u0440u0430u0441u0441u0442u043eu00a0u043du0438u0439, u0432u0437u00a0u0442u044bu0445 that familiar, groups of "winners".and on the basis of u0441u0440u0430u0432u043du0435u043du0438u00a0 amount with a threshold of u043eu043fu0440u0435u0434u0435u043bu00a0u044eu0442 class testing u0438u0437u043eu0431u0440u0430u0436u0435u043du0438u00a0, the u0444u043eu043bu043bu0438u043au0443u043bu00a0u0440u043du043eu0439 u0430u0434u0435u043du043eu043cu0435 or u0444u043eu043bu043bu0438u043au0443u043bu00a0u0440u043du043eu043cu0443 cancer what prominent gland.