diff --git a/01_dataset.ipynb b/01_dataset.ipynb index 3edb2f2..0bfa4ce 100644 --- a/01_dataset.ipynb +++ b/01_dataset.ipynb @@ -48,8 +48,8 @@ }, "outputs": [], "source": [ - "img_path = r\"data/dataset_old/img/yangeng.bmp\"\n", - "label_path = r\"data/dataset_old/label/yangeng.bmp\"\n", + "img_path = r\"data/dataset/img/85.bmp\"\n", + "label_path = r\"data/dataset/label/85.bmp\"\n", "# 读取图片和色彩空间转换\n", "img = cv2.imread(img_path)\n", "label_img = cv2.imread(label_path)\n", @@ -62,7 +62,7 @@ }, { "cell_type": "code", - "execution_count": 28, + "execution_count": 3, "metadata": { "pycharm": { "name": "#%%\n" @@ -77,7 +77,7 @@ }, { "cell_type": "code", - "execution_count": 29, + "execution_count": 4, "metadata": { "pycharm": { "name": "#%%\n" @@ -86,10 +86,8 @@ "outputs": [ { "data": { - "image/png": "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\n", - "text/plain": [ - "
" - ] + "text/plain": "
", + "image/png": "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\n" }, "metadata": { "needs_background": "light" diff --git a/02_classification.ipynb b/02_classification.ipynb index e65d8ed..6fd34e3 100644 --- a/02_classification.ipynb +++ b/02_classification.ipynb @@ -12,7 +12,7 @@ }, { "cell_type": "code", - "execution_count": 3, + "execution_count": 7, "outputs": [], "source": [ "import numpy as np\n", @@ -30,7 +30,7 @@ }, { "cell_type": "code", - "execution_count": 4, + "execution_count": 8, "outputs": [], "source": [ "train_from_existed = False # 是否从现有数据训练，如果是的话，那就从dataset_file训练，否则就用data_dir里头的数据\n", @@ -70,7 +70,7 @@ }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 9, "outputs": [], "source": [ "dataset = read_labeled_img(data_dir, color_dict=color_dict, is_ps_color_space=False)\n", @@ -99,7 +99,7 @@ }, { "cell_type": "code", - "execution_count": 6, + "execution_count": 10, "outputs": [], "source": [ "if len(dataset) > 1:\n", @@ -130,7 +130,7 @@ }, { "cell_type": "code", - "execution_count": 7, + "execution_count": 11, "outputs": [], "source": [ "# 对数据进行预处理\n", @@ -147,13 +147,13 @@ }, { "cell_type": "code", - "execution_count": 8, + "execution_count": 12, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ - "38411it [02:01, 316.59it/s] \n" + "31173it [01:21, 380.30it/s] \n" ] }, { @@ -162,12 +162,12 @@ "text": [ " precision recall f1-score support\n", "\n", - " 0 1.00 1.00 1.00 11523\n", - " 1 1.00 1.00 1.00 9603\n", + " 0 1.00 1.00 1.00 9352\n", + " 1 1.00 1.00 1.00 7793\n", "\n", - " accuracy 1.00 21126\n", - " macro avg 1.00 1.00 1.00 21126\n", - "weighted avg 1.00 1.00 1.00 21126\n", + " accuracy 1.00 17145\n", + " macro avg 1.00 1.00 1.00 17145\n", + "weighted avg 1.00 1.00 1.00 17145\n", "\n" ] } diff --git a/config.py b/config.py index 0b2782e..2147a2b 100644 --- a/config.py +++ b/config.py @@ -24,10 +24,10 @@ class Config: spec_size_threshold = 3 # rgb模型参数 - rgb_tobacco_model_path = r"models/tobacco_dt_2022-07-26_15-57.model" - rgb_background_model_path = r"models/background_dt_2022-07-27_08-11.model" + rgb_tobacco_model_path = r"models/tobacco_dt_2022-08-05_10-38.model" + rgb_background_model_path = r"models/background_dt_2022-08-05_10-41.model" threshold_low, threshold_high = 10, 230 - threshold_s = 175 + threshold_s = 190 rgb_size_threshold = 4 # mask parameter @@ -35,4 +35,3 @@ class Config: # save part offset_vertical = 0 - diff --git a/main_test.py b/main_test.py index c682d57..b244d92 100644 --- a/main_test.py +++ b/main_test.py @@ -175,5 +175,5 @@ if __name__ == '__main__': parser = argparse.ArgumentParser(description='Run image test or ') tester = TestMain() - tester.pony_run(test_path=r'E:\zhouchao\8.4\yangeng', - test_rgb=True, test_spectra=True, get_delta=False, convert=True) + tester.pony_run(test_path=r'E:\zhouchao\8.4\zazhi', + test_rgb=True, test_spectra=True, get_delta=False, convert=False)