What is Ikena?

Ikena is a video enhancement software application featuring MotionDSP’s patent-pending super-resolution technology. Able to run on any Windows PC (laptop or desktop), its powerful multi-frame enhancement technology can quickly and significantly improve poor video from less-than-ideal sources, such as consumer mobile phones, digital cameras, or surveillance cameras

What are the system requirements for Ikena?

For Ikena

• Windows 7 64 bit; Windows XP 32bit , Service Pack 3.
• CPU – Intel  or AMD dual-core processor, 2 GHz or higher (SSE-2 compatible required). Quad-core and higher recommended for even better performance.
• RAM – 2GB
• Hard Drive – 50MB available space
• USB – 1 port available for activation dongle
• Matrox video cards are not supported

For Ikena GPU/Ikena ISR

• Windows 7 64 bit; Windows XP 32bit , Service Pack 3
• NVIDIA Quadro/Tesla Driver version 197 or higher
• CPU – Intel or AMD dual-core processor, 2.5 GHz or higher (SSE-2 compatible required). Quad or 6-core CPU recommended. 
• RAM – 4GB
• Hard Drive – 50MB available space
• USB – 1 port available for activation dongle
• CUDA GPU compute card
  • Mobile: Quadro 3800M, Quadro 4000M, Quadro 5000M
  • Workstation/Server: NVIDIA Quadro 4000, 5000, 6000 or NVIDIA Tesla C2050 (GPU compute) + NVIDIA NVS 290 (for display)
• Video recording input (optional): DV/1394-based video input with the Canopus ADVC 110
• Matrox video cards are not supported

What file formats can Ikena import and export?

Ikena supports several input file formats, video codecs, and audio codecs, including the following:

• File formats: DV, Flash, Real, WindowMedia, Quicktime, AVI, 3GP, 3G2, MP4
• Video Codecs: Uncompressed RGB, H264, MPEG-4, H263, WMV7, WMV8, WMV9, VC-1
• Audio Codecs: MPEG, AC3, AAC, WMA, ADPCM, AMR

Ikena can export both high-res still images, and high-res video.

High-res stills can be exported in uncompressed BMP, compressed JPEG, PNG and TIFF formats. High-res video can be exported in uncompressed AVI, Motion-JPEG-compressed AVI, XVID-compressed AVI, Digital Video DV25 and Windows Media Video WMV.

How is Ikena's technology different than competitive programs which use Frame Averaging?

First-generation video forensic software packages were designed to be used by highly-skilled video forensic technicians. Ikena was designed to be automated, fast, and usable with very little training.

Ikena's patented super-resolution algorithms differ from Frame Averaging (also known as Integration) in several ways:

1. Frame averaging doesn't increase resolution (i.e., the resolution of your output is no higher resolution than your input resolution).
2. Frame averaging can't deal with any motion (as it just averages the pixels at the same spatial positions in all frames). If there is any movement between frames, frame averaging falls apart.
3. Frame averaging requires manual "de-warping" of video frames. If there has been any geometric distortion between the frames (for example, if the camera or object angle changes slightly), each frame needs to be "de-warped" manually. Manual de-warping is possible using fairly simple methods (translation, rotation, or zoom) and it may work only for rigid objects/areas. MotionDSP's super-resolution provides for sophisticated, and automatic de-warping of frames, without any manual steps. 
4. In addition to points 1-3, super-resolution is superior to frame averaging because it uses much more sophisticated "pixel combination model". -- that is, super-resolution selects the best pixels for re-construction.

For example, in Frame Averaging, when you frame average 10 frames, the math is:

• Output = (Input1 + Input2 + ... + Input10)/10
• Super-resolution, uses a statistically and content-optimized function of the 10 input values
• Output = complex_function {Input1, Input2, ...., Input10}
  

Ikena for US Federal customers

GSA Advantage

To better serve our US federal customers, MotionDSP has made its products available on the GSA schedule. Please contact MotionDSP sales for more information on purchasing Ikena and Ikena ISR off the GSA.

What are GSA Schedules?

GSA stands for General Services Administrative Contract. GSA Schedules are contracts awarded by GSA's Federal Supply Service to make it easier for federal agencies to find and purchase the products, services and equipment they need. GSA has already negotiated fair and reasonable prices, and places all of these options in a single location which simplifies the buying process.

What are the basic features of Ikena?

Ikena provides:

• powerful video and single-frame enhancement – Ikena's set of processing filters allows users to remove noise and compression artifacts and increase video resolution by factors up to 9x (3x the pixels in both the vertical and horizontal axis). MotionDSP’s technology extracts information from multiple frames of video, revealing details not visible in the original sequence.
• still image capture - capture enhanced, high-resolution stills images in one click.
• video playback - for all of the commonly used video file formats
• export of results – the enhanced video and enhanced still images can be exported in various video and image formats.

How do you import video into Ikena?

Import video from local drive

Choose File -> Import video from: Disk or click the the Disk button in the project bar. Ikena supports most commonly used video formats - AVI, MOV, WMV, 3GP, 3G2, and many more. File information on the imported video file (file size, duration and frames per second) is shown in video info box just above the video.

Open video from an Internet link

Ikena can download videos from video sharing services, such as YouTube. Choose File -> Import video from: Link. Paste the YouTube video URL into the Download Video dialog. After starting the download, choose a filename to save the downloaded video as. After the download is finished, the downloaded video will be displayed in the Project list.

Rename video

Choose Project -> Rename video in the main menu, video file name in video info box, or right click video file name. This enabled editing of video name. Press Enter to complete editing and rename video file. NOTE: This option renames original video file. If video was opened from a read-only source (such as CD, DVD or network storage) rename will fail.

How do you import a sequence of still images into Ikena?

1. To import an image sequence into Ikena, open the still image file you want to use as its first frame (typically the first frame in the the folder). No special treatment of image sequences is required (e.g., just draging-and-dropping an image file into Ikena works).
2. All image files must have the same resolution.

3. The files need to be named sequentially (ie: when you view them sorted by name in Windows, they should arrange themselves in the correct order). The part of the filename that comes before the first digit is interpreted by Ikena as the "root" string. For example with a file called "video_sequence_0001.png", the root string is "video_sequence_"

   The images are then sorted by the name and added to Ikena. Note: the numbers don't have to be consecutive -- that is, there can be missing numbers, for example video_01, video_02, video_04, video_06 will work.

   To review:
   • video_sequence_0002.png
   • video_sequence_0010.png
   will be treated as a part of image sequence when "video_sequence_0001.png " is opened, but
   • video_seq_0003.png or
   • video_sequence_0004.jpg
   will not.
4. The still image formats supported are png, jpg, and bmp.

Does Ikena have presets?

Yes. To begin, choose a preset that best suits the video you're working with: Ikena ships with a number of presets already created, and you can save new presets by choosing Edit -> Save Preset. Once you save them they will show up in the drop-down menu. See the manual for how to create your own presets.

How do you process a still frame with Ikena?

Ikena has the option to process only a selected region of the original image.

1. First, find the frame with information of your interest on the progress bar.
2. Click the ROI button or hold the CTRL key to put the mouse cursor in drawing mode.
3. Using the mouse, drag a box around the area of interest you want to enhance.
4. Turn on the live preview mode.
5. Start processing (click Process frame button or F5).

The resulting still image will have your region of interest enhanced, surrounded by an "non-enhanced" version of the original frame.

To get the best results, when you are creating a still image, experiment with:

• the area of the image you select - try slightly different areas for better results,
• the number of matching frames (see the section on matching frames in the "Advanced Processing" section),
• the specific time point you wish to enhance - move the cursor slightly ahead or back - a different moment might get a better result.

How do you enhance a portion of a video in Ikena?

Ikena lets you limit your enhancement to a specific time segment of the video - this speeds up processing by only enhancing the section of the video you need. Using the trim points, you can select start and end points. To set the section start and end points, pause the video where you want processing to start, and click on the Set section start button. Then, pause the video at the point where you want processing to stop, and click the Set section end button. A blue line will indicate your trimming selection. Using the Process Video button or F6 you will enhance the selected part of the video.

If you want to process the entire video or to select a different segment, use the Clear section button.

How do you set the Resolution Enhancement parameters?

Use "Auto Settings"

Ikena has a sophisticated "Auto Settings" feature that automatically scans your video and chooses appropriate enhancement parameters. This should be your first step. Then, you can tweak from there.

Matching Frames

Some basic rules of thumb for matching frames are:

• For low frame-rate video like mobile phones video (5-10fps), or for videos with high motion, use 6-9 matching frames.
• For medium frame-rate video like that from digital cameras (15fps), or for videos with moderate motion, use 10-12 matching frames.
• For high frame-rate video, like that from a DV camcorder or security camera, or for video with low motion, you can use 15 for matching frames.
• For de-interlacing, start with low matching levels - like 2 or 3, as the de-interlacing process effectively doubles the number of matching frames.

Fusion Level

In most cases, leaving the Fusion Level setting at the default value of 0 will yield good results, however, in video forensic-type applications where you know specifically what you are trying to enhance, setting this parameter manually can be beneficial. In particular, if you have a video which has been captured from a distance, and the area you want to enhance is part of the background scene - for example, a license plate on a parked car, then you can set Fusion Level to a high positive value (5 - 10). This is particularly powerful when you combine it with using a Region of Interest selection (ROI), as you can focus your processing on exactly what you want to enhance, then set the fusion level specific to the object or area you're trying to enhance.

Super Resolution

The super-resolution factor sets the amount of enlargement or scaling of the video. It works as a square of the super-resolution factor, that is, an factor of "2" gives you 4 times the number of pixels - 2x the number of pixels horizontally, and 2x the pixels vertically. The software will produce an image or video at whatever factor you set, however, getting actual "results" from a large setting depends on the content of your video. Also, setting the super-resolution factor to a higher factor squares the processing time required, as the number of pixels that must be processed is so much higher. A rule of thumb is that in order to get a super-resolution factor of 2 with visible results, you need at least 5-6 frames of information, with 9-15 being preferable.

How do you use Contrast Enhancement?

Ikena has an advanced algorithm for contrast-limited adaptive histogram equalization. This process uses both automatic contrast adjustment, as well as a proprietary algorithm that combines exposure information from multiple frames of information to create an output image that has the image quality comparable to images created with a longer exposure time. This has the effect of both suppressing noise, and revealing new detail within the image/video.

Level You can change the amount of contrast enhancement by moving the "Contrast Level" slider from 0 (no change in contrast) to 100 (maximum contrast enhancement). The default value for this parameter is 30.

Contrast setting.

What video filters does Ikena have?

Ikena implements the following video/image filters:

• Super-resolution: Ikena uses super-resolution reconstruction to re-create each frame of video with the best information found from neighboring frames. This results in better detail and significantly reduced noise and artifacts. Ikena's super-resolution automatically does highly accurate motion estimation, registration, super-resolution reconstruction, and fusion.
• Stabilization: Ikena implements state-of-the-art stabilization.
• Deblurring: Ikena uses bilateral total variation deblurring for high-quality deblurring with adjustable focus (kernel size), selectable kernels (gauss, disk), and noise reduction.
• Contrast: Ikena implements MotionDSP's proprietary histogram equalization, which dynamically stretches lighting and contrast to adjust for poorly lit scenes, with adjustable sensitivity.
• Dynamic Light and Color: Ikena implements dynamic autolevels of shadows and highlights, allowing Ikena to adapt to lighting changes in content dynamically -- for example, it will brighten an indoor sequence, and if the camera moves outdoors in the same sequence, it will adjust the lighting to adjust for the outdoor brightness.
• Shadows and Highlights: Ikena allows manual adjustment of shadows and highlights
• Gamma: users can manually adjust gamma
• Color Saturation: users can manually adjust color saturation, from black and white to fully saturated
• Color: users can also adjust color balance between warm, neutral, and cold

What is the difference between Ikena and Ikena GPU

Ikena and Ikena GPU are identical in their basic feature sets.

What sets Ikena GPU apart from Ikena is speed. Ikena GPU can utilize OpenCL or CUDA-compatible graphics cards from AMD and NVIDIA, and with a high-end card like the AMD FirePro(TM) V8800, Ikena GPU can process video up to 6 times faster than a high-end Intel CPU. We've used a 6-core Intel CPU below. For 2 and 4-core CPUs, the difference is even greater -- up to 15x faster.

In addition, Ikena GPU comes with our Evidence Spotlight tool included in the purchase price, a $1000 value.

Specific performance numbers below.

Live Preview Processing (Frames per second)

Video Resolution	CPU Intel 6-core	GPU AMD FirePro V8800	GPU NVIDIA GTX 480 GPU
480p (1x SR)	24	30 (max)	30 (max)
480p (2x SR)	6	30 (max)	30 (max)
720p (1x SR)	10	30 (max)	30 (max)
1080p (1x SR)	4.5	30 (max)	18

Processing Parameters:

• 7 matching frames (slow option)
• Stabilization (120%)
• Deblurring (Level 40, Focus 0.40)
• Contrast (30)
• Light and Contrast (auto levels on)

Can Ikena Process HD Video?

Yes, Ikena, Ikena GPU, and Ikena ISR can all process HD video, up to 1080P (1920x1080).

Performance (speed of processing) will depend on the PC hardware used, and the settings used in Ikena.

We recommend the highest powered GPUs for HD processing (such as the NVIDIA Tesla C2050 GPU).

Due to OS memory limitations, some filters, such as MotionDSP's super-resolution algorithm (called "resolution" in Ikena) will be limited to certain SR zoom factors, but this can be mitigated with the use of the "region of interest" control, which limits processing to a certain area of the screen.

What is the difference between Ikena and vReveal?

vReveal is a consumer product from MotionDSP which is designed to make consumer videos "look great" -- improving the subjective quality of the video, but it is not designed, licensed, or recommended for any forensic use.

vReveal does not use MotionDSP's patented, automated super-resolution technology, and thus, vReveal won't help you reconstruct and read license plates from security videos, or grab high-res images of faces from video. For that, you need Ikena.

Ikena has been designed from the ground up for video enhancement for law enforcement and forensics labs, with powerful tools for capturing, enhancing and presenting court-admissible evidence. It uses MotionDSP's patented, automated super-resolution algorithms to reconstruct each frame of video with the best information from surrounding frames of video, objectively increasing resolution and detail while significantly reducing noise. It was developed with direct input and requirements from several of the world's leading video forensic labs. Results from Ikena have been used to prosecute court cases in both the US and the UK.

For more details about Ikena, please visit www.motiondsp.com.