Solid-state micropores can measure distinct mechanical properties of the critical diseased cells at its finer granularity and can drastically improve the prognosis and treatment of a disease at its early stages. However, these devices suffer from huge amount of raw data, complicated with the inherit sensor noise and baseline artifacts. Furthermore, the stateof- The-art detection and analysis involves visual inspection, and the available softwares can only analyze a subset of the acquired data, which makes the overall process timeconsuming, tedious, and error-prone. Therefore, it is important to improve and automate the data processing strategy in order to effectively detect and identify the diseased cells in the raw data. In this paper, we present a pattern detection approach based on the moving-average filtering technique that smooths the raw data to get rid of baseline wanders and reduce the noise. Our proposed technique computes the threshold against the electric pulses using the mean and standard deviation of the smoothed data in order to detect pulse features including the width and amplitude. Our proposed framework detects cancer cells with 63% accuracy in a mixture of cancer cells, Red Blood Cells (RBCs), and White Blood Cells (WBCs), and can process 10 gigabytes of raw data, collected from the translocation of 0.5 milliliter of blood sample via micropore, in about 6 minutes. Copyright © (2014) by the International Society for Computers and Their Applications.