Ad Insertion

[ultimate_heading main_heading=”PRODUCT” spacer=”line_only” spacer_position=”bottom” line_height=”1″ main_heading_font_size=”desktop:36px;” main_heading_line_height=”desktop:46px;”][/ultimate_heading]
[ult_tab_element tab_style=”Style_4″ tab_animation=”None” tab_background_color=”#eaeaea” tab_hover_background_color=”#c9060d” acttab_background=”#c9060d” tabs_border_radius=”0″ disp_icon=”Disables” resp_style=”Title”][single_tab title=”Explanation” tab_id=”1562610615228-4-6″]

The objective of Video Advertisement Placement system is to insert advertisements into videos by analysing the contents of the video, determining the best locations and time frames to insert the advertisements, selecting advertisements related to the contents of the video, and placing them in the video by applying necessary transformations in 3D. This system provides a way to show advertisements without obstructing important parts of the video and thus without disturbing the viewers. Advertisements can be placed into videos as if they were physically there when the videos were recorded. For this purpose, state-of-the-art machine learning, computer vision, and video processing methods are utilized.

Used Technologies:
• Semantic / content-based / multimodal video analysis (using mostly deep learning)
• Image tagging & video classification
• Object recognition, image & video captioning
• Face & text detection
• Saliency detection
• Speech recognition & text analysis
• Automatically find the best spatial location and time
• Transforming the ad in 3D accordingly
• Homography matrix estimation
• Keeping the ad on the surface and handling occlusions
• SIFT keypoints, object trackers, background subtraction
• Content-based ad selection from database

[/single_tab][single_tab title=”Unique Advantages” tab_id=”1562610767710-3-6″]

Objective is to develop a system for showing ads in videos without disturbing the viewers by identifying the best location and time frame in a given video, selecting semantically related advertisements from a database, and placing them properly in the video.

[/single_tab][single_tab title=”Application Scenarios” tab_id=”1561669169545-3-9″]

• Scenario 1 : Automatic License Blurring
• Scenario 2 : Yolo Human Detection
• Scenario 3 : Pose Estimation
• Scenario 4 : Optical Character Recognition

[/single_tab][/ult_tab_element]
[ultimate_heading main_heading=”PRODUCT VISUALS” spacer=”line_only” spacer_position=”bottom” line_height=”1″ main_heading_font_size=”desktop:36px;” main_heading_line_height=”desktop:46px;”][/ultimate_heading]
[ultimate_heading main_heading=”RESOURCES” spacer=”line_only” spacer_position=”bottom” line_height=”1″ main_heading_font_size=”desktop:36px;” main_heading_line_height=”desktop:46px;”][/ultimate_heading]
[icon_counter flip_box_style=”advanced” box_border_style=”solid” box_border_size=”2″ icon=”Defaults-play-circle-o” icon_size=”40″ icon_color=”#c9060d” block_title_front=”Global Link” block_desc_color=”#000000″ block_front_color=”#ffffff” box_border_color=”#a4a4a4″ block_title_back=”Global Link” block_back_text_color=”#ffffff” block_back_desc_color=”#ffffff” block_back_color=”#c9060d” box_border_color_back=”#a4a4a4″ custom_link=”1″ button_link=”||target:%20_blank|” button_text=”Go to Page” button_txt=”#ffffff”]