+ Add docs for newly-added OP image_face_count_filter

configs/config_all.yaml

@@ -311,6 +311,10 @@ process:
       min_ratio: 0.333                                        # the min aspect ratio of filter range
       max_ratio: 3.0                                          # the max aspect ratio of filter range
       any_or_all: any                                         # keep this sample when any/all images meet the filter condition
+  - image_face_count_filter:                                # filter samples according to the face count in images
+      cv_classifier: ''                                       # OpenCV classifier path for face detection. By default, we will use 'haarcascade_frontalface_alt.xml'.
+      min_face_count: 1                                       # the minimum number of faces required for samples.
+      max_face_count: 1                                       # the maximum number of faces required for samples.
   - image_face_ratio_filter:                                # filter samples according to the face area ratios in images (r=face_area/image_area). If multiple faces are available, we use the largest one.
       cv_classifier: ''                                       # OpenCV classifier path for face detection. By default, we will use 'haarcascade_frontalface_alt.xml'.
       min_ratio: 0.0                                          # the min face area ratio of filter range

data_juicer/ops/filter/__init__.py

@@ -3,17 +3,17 @@
                audio_nmf_snr_filter, audio_size_filter,
                average_line_length_filter, character_repetition_filter,
                flagged_words_filter, image_aesthetics_filter,
-               image_aspect_ratio_filter, image_face_ratio_filter,
-               image_nsfw_filter, image_pair_similarity_filter,
-               image_shape_filter, image_size_filter,
-               image_text_matching_filter, image_text_similarity_filter,
-               image_watermark_filter, language_id_score_filter,
-               maximum_line_length_filter, perplexity_filter,
-               phrase_grounding_recall_filter, special_characters_filter,
-               specified_field_filter, specified_numeric_field_filter,
-               stopwords_filter, suffix_filter, text_action_filter,
-               text_entity_dependency_filter, text_length_filter,
-               token_num_filter, video_aesthetics_filter,
+               image_aspect_ratio_filter, image_face_count_filter,
+               image_face_ratio_filter, image_nsfw_filter,
+               image_pair_similarity_filter, image_shape_filter,
+               image_size_filter, image_text_matching_filter,
+               image_text_similarity_filter, image_watermark_filter,
+               language_id_score_filter, maximum_line_length_filter,
+               perplexity_filter, phrase_grounding_recall_filter,
+               special_characters_filter, specified_field_filter,
+               specified_numeric_field_filter, stopwords_filter, suffix_filter,
+               text_action_filter, text_entity_dependency_filter,
+               text_length_filter, token_num_filter, video_aesthetics_filter,
                video_aspect_ratio_filter, video_duration_filter,
                video_frames_text_similarity_filter, video_motion_score_filter,
                video_nsfw_filter, video_ocr_area_ratio_filter,
@@ -29,6 +29,7 @@
 from .flagged_words_filter import FlaggedWordFilter
 from .image_aesthetics_filter import ImageAestheticsFilter
 from .image_aspect_ratio_filter import ImageAspectRatioFilter
+from .image_face_count_filter import ImageFaceCountFilter
 from .image_face_ratio_filter import ImageFaceRatioFilter
 from .image_nsfw_filter import ImageNSFWFilter
 from .image_pair_similarity_filter import ImagePairSimilarityFilter
@@ -102,6 +103,7 @@
     'ImageSizeFilter',
     'VideoWatermarkFilter',
     'WordsNumFilter',
+    'ImageFaceCountFilter',
     'ImageFaceRatioFilter',
     'FlaggedWordFilter',
     'WordRepetitionFilter',